Population Differentiation in the Dingo: Biogeography and Molecular Ecology of the Australian Native Dog Using Maternal, Paternal and Autosomal Genetic Markers

Population differentiation in the dingo: biogeography and molecular ecology of the Australian Native Dog using maternal, paternal and autosomal genetic markers

Kylie Mae Cairns

A thesis submitted for the degree of Doctor of Philosophy in the Faculty of Science

School of Biotechnology and Biomolecular Sciences The University of New South Wales, Sydney, Australia

2014

I would like to dedicate this thesis in memory of A/Prof Alan Wilton for inspiring my

interest in and appreciation of dingoes.

Originality Statement

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorize University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.’

Kylie Mae Cairns

28 August, 2014

‘I hereby declare that this submission is my own work, and to the best of my knowledge contains no materials previously published or written by other persons, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to this research by others with whom I have worked at UNSW or elsewhere is explicitly acknowledged in the thesis. I also declare that all intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception, or in style, presentation and linguistic expression is acknowledged.’

Kylie Mae Cairns

28 August 2014

Authenticity Statement

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Kylie Mae Cairns

28 August, 2014

Abstract

Dingoes are an essential feature of the Australian landscape, a trophic regulator and apex carnivore. As an ancient dog they may also hold the key to investigating the evolution, history and origin of domestic dogs. Conservation efforts are hampered by a lack of consistency in federal and state legislation. In the wild dingoes are threatened by hybridisation with domestic dogs. Conservation groups have posited that different ecotypes of dingo exist, however little biogeographical data has been collected.

This thesis explores the phylogeny, biogeography and genetic identity of the dingo using maternal, paternal, autosomal and whole genome genetic markers. Previous genetic studies lacked a broad geographic sample inhibiting the observation of broad biogeographical patterns. Thus, for this research, dingoes from five broad geographic populations across the entire Australian continent were interrogated; The Kimberley

(Western Australia), The Gibson Desert (Western Australia), The Simpson Desert

(Northern Territory), Fraser Island (Queensland) and The Australian Alpine region

(New South Wales/Victoria/Australian Capital Territory). New Guinea Singing Dogs, the wild dog of Papua New Guinea, were also incorporated as the closest genetic relative to the dingo.

The main finding of this thesis research is that there are at least three genetically distinct geographically subdivided populations of dingo; southeastern, Fraser Island and northwestern. Genetic data suggests that the dingo lineages diverged outside Australia.

Mitochondrial and Y chromosome evidence further suggests that these populations may

I be the result of multiple introductions into Australia. The New Guinea Singing Dog was observed to have shared ancestry with the dingo. Paternal introgression from domestic dogs was observed, particularly in southeastern Australia. Additionally strong inbreeding was observed in the Fraser Island dingo population. These findings have significant implications for the management and conservation of dingoes. First, the three distinct dingo populations should be managed separately, with limited mixing in captivity and restricted human-mediated translocations in the wild. Second, hybridisation with domestic dogs is a particular threat to the southeastern dingo population and lethal control methods may be increasing male biased introgression into dingoes. Third, inbreeding possibly compromises Fraser Island dingoes; further genetic surveys are imperative and current management strategies may be unsustainable.

Acknowledgements

Writing a thesis, particularly a doctorate, is a long journey. Like raising a child, “it takes a village”. I would like to thank the various people who have helped me through this journey from both my professional and personal life. I would not be the scientist (or person) I am today without you.

Firstly, I must acknowledge the thoughtful guidance and support of my supervisor Bill

Ballard. I thank you for your scientific tutelage and advice, you have taught me many valuable lessons about science and academia. Thank you for encouraging me to take on a PhD and guiding me through it.

To my co-supervisor A/Professor Alan Wilton, who sadly passed away in 2011 during my candidature. Thank you for teaching me all that you could about dingoes, dogs and genetics. You are greatly missed by friends, family and colleagues.

To my co-supervisor Dr Paul Waters, who kindly stepped in to fill a void in 2012 following A/Prof Wilton’s passing. I am indebted to you for your excellent continued help, advice and kindness during the end of my tenure.

To my thesis committee, Professor Mike Archer and Dr Mike Letnic at UNSW, I thank you for your thoughtful and useful input into my research project. Many thanks for always making the time for me.

III

To my dear friend Dr. Barbara Zangerl, I wish we had met under happier circumstances.

Thank you for the guidance regarding microsatellites, SNP genotyping, canine genetics and science in general. I am indebted to you as well for reading various drafts of my thesis and giving valuable constructive criticism.

To all the members of the Ballard Lab throughout my candidature: the post-docs Nadia

Urosevic, Louise Puslednik, Richard Melvin, Jonci Wolff, Nicolas Pichaud and Martin

Horan; the graduate students Theodore Orfanos, Olaf Bressel, Eric Ngo, Michael

Nafisinia, Marie Messmer, Melina Chok, Wen Aw, Rijan Bajracharya , Mihi Hwang and Cha Yeon Lee and to the lab assistants Preetha Sujit, Lize Toman, Mikey Amiezer and Jason Wang. Thank you for being such a friendly, helpful and welcoming crowd.

Especially for the many times you offered me advice concerning wet lab work, listened to lab presentations about dingoes or read my paper drafts. Particularly to Mart and

Jonci thank you for teaching me about PCR – troubleshooting, numts, hot start etc, it was invaluable. It was a pleasure to interact with each and every one of you.

I wish to give special acknowledgement to Pann Pann Chung and Carolina Correa, we started our PhDs ‘together’ and now we have all ‘finished’. Thank you both for the continued support, willingness to help with editing, reading, commenting on my work, discussions about women in science and most importantly friendship.

To my family, Mom, Dad, and my siblings Hilary, Alastair and Tim. Thank you for supporting me through this journey and willingly listening to endless conversations about dingoes. Thanks must especially go to my Father, Iver, for countless discussions

IV about academia, science and peer-review. My mother Ann, sister Hilary and mother-in- law Janine; without your assistance in caring for Casper I could not have finished, thank you.

To my husband Nathan, I must thank you for the unfailing support and positivity.

You’ve been with me every step of this journey; listening to endless talks about dingoes, science, grant writing and offering advice with patience. Thank you for putting up with my absentmindedness, passion for science and supporting me to undertake a career that will never be 9-5. To my son Casper, you make it all worthwhile. It is hard spending time away from you, even harder given you are so young, however I had to do this. I hope you are proud of me when you are older and the sacrifice was worth it.

In addition, I would like to acknowledge the contributions of the various dingo trappers, scientists, National Parks rangers, citizens and conservation groups whom have sent Dr.

Alan Wilton (and myself) dingo samples. In particular I must thank Dr. Danielle

Stephens whom kindly offered me the use of dingo samples from her PhD research and was always available to discuss the intricacies of dingo genetics. Without these contributed samples this research would not have been possible.

Academic Acknowledgements

Chapter 1

The authors would like to acknowledge the Ballard/Wilton laboratory group and

Michelle Potter at the University of New South Wales for comments on the manuscript.

We thank Dr. Denise Donlon (University of Sydney) for providing permission to reproduce the photograph in Figure 1-1 and providing access to the J. L. Shellshear

Museum archives (University of Sydney). We thank H.C. Cairns for permission to reproduce the photographs in Figure 1-2 from his book, co-authored with Bill

Yidumduma Harney, Dark Sparklers and providing insight into Indigenous Australian culture and spiritualism concerning the dingo. The dingo 454 sequence reads were generated at Pennsylvania State University by Stephan Schuster. The dingo 454 sequence reads were aligned to the dog reference genome (v2.1) by Web Miller

(Pennsylvania State University, USA) and Sven Warris (Hanze University of Applied

Sciences, The Netherlands). Sven Warris also offered comments on the manuscript.

Thanks to Peter Savolainen (Royal Institute of Technology, Sweden) for providing information to the authors concerning unpublished research on Y chromosome haplotypes in dingoes. We thank Britt-Louise Carlsson (University of New South

Wales) for unpublished data on indels and SNP variation in dingoes. The authors would like to thank two reviewers for comments, which have improved this chapter.

Chapter 2

Appreciation goes to the late author Alan N. Wilton whom sadly was unable to review the manuscript. Gratitude to Matthew Wong (UNSW) for preliminary mitochondrial data and primers collected during his honours thesis (unpublished). Danielle Stephens

(UWA) provided a subset of pre-screened dingo samples for this study. Thanks to Peter

Savolainen (KTH Biotechnology) for providing the Indonesian dog samples and Mike

Archer (UNSW) and Mike Letnic (UNSW) for comments. The authors appreciate advice concerning phylogenetic analyses and divergence dating intricacies from Simon

Ho (USYD). This manuscript was improved by comments from the Ballard and Wilton lab groups, Dr. Barbara Zangerl (UNSW), Dr G. Larson (Durham University) and anonymous reviewers. This work was supported by a Hermon Slade Foundation

Research grant (HSF11/6 to JWOB, ANW and KMC).

Chapter 3

Thanks must be directed to Dr Paul Waters (UNSW), Dr Carolina Correa (UNSW) and

Dr Barbara Zangerl (UNSW) for providing valuable and constructive comments on the initial manuscript. Dr Danielle Stephens and Dr Alan Wilton provided access to samples for this study. As well appreciation must go to the large number of conservation organisations, dingo trappers, land managers and government agencies for contributing dingo samples to genetic research projects on the dingo. Janice Koler-Matznick and other NGSD conservationists in North America supplied the NGSD samples for this project.

VII

Chapter 4

Acknowledgements should be directed to Dr Danielle Stephens for sharing dingo samples used in this research. Dr Barbara Zangerl provided guidance concerning the analysis of SNP genotype data. Dr Alan Wilton provided the original idea and impetus to using SNP data in dingoes; sadly he did not get to see the end result.

Grant and Scholarship Funding

I would like to thank the Hermon Slade Foundation for supporting my PhD research through a research grant awarded to Bill Ballard, Alan Wilton and myself in 2011

(HSF11/06). Additionally I could not have undertaken a PhD without an APA scholarship from the Australian government and I would not have finished without a maternity leave scholarship from the UNSW Faculty of Science.

VIII

Notes on Thesis Structure

Chapter 1, the introduction to this thesis, has been published as a peer-reviewed book chapter. Although only a single chapter has been published from this thesis the rest has been prepared as a series of papers. As such, there is some repetition between chapters so that each is independently publishable. Genetic sequence data collected in this thesis has not been publicly released, however data will be placed in a public repository such as GenBank upon publication.

All chapters with the exception of Chapter 5 have been prepared as co-authored publications, already published or to be published after submission of this thesis. Data in Chapter 3 was collected as part of collaboration with Ben Sacks and Sarah Brown from University of California Davis. Chapter 4 contains data collected as part of collaboration with Adam Boyko from Cornell University and Janice Koler-Matznick from the NGSD Conservation Society.

Chapter 1 was written in 2010 at the beginning of my candidacy as such there is some inconsistency between the thesis aims and literature reviewed in this chapter and the following thesis chapters.

Author Contributions

Chapter 1: Cairns KM, Wilton AN and Ballard JWO

Contributions: Primary manuscript text written by KMC with a small contribution, approximately 4 paragraphs as well as Table 1-2 and Figure 1-4, from ANW regarding

SNP genotyping. JWOB and ANW provided comments and editing to the draft.

(Published 2011)

Note: This chapter was reviewed by Dr. Mike Letnic, School of Natural Sciences,

University of Western Sydney, Australia and Dr. Luke K-P Leung, School of Animal

Studies, University of Queensland, Australia.

Chapter 2: Cairns KM, Wilton AN and Ballard JWO

Contributions: Data collection, analysis and manuscript text completed by KMC. ANW and JWOB provided advice and comments concerning experimental design and analysis. JWOB also provided constructive criticism and editing to the manuscript.

Sadly ANW passed away before the manuscript was completed.

Chapter 3: Cairns KM, Sacks BN, Brown SK. and Ballard JWO

Contributions: Mitochondrial DNA data collection, analysis and manuscript text completed by KMC. BNS and SKB provided advice concerning experimental design,

X analysis and collected the Y chromosome data. JWOB provided advice and comments concerning experimental design.

Chapter 4: Cairns KM, Boyko AR, Koler-Matznick J and Ballard JWO

Contributions: DNA extraction, analysis and manuscript text completed by KMC. ARB provided advice concerning experimental design, analysis and collected the SNP genotype data. ARB and JKM contributed NGSD SNP genotype data to this project.

JWOB provided advice and comments concerning experimental design.

Chapter 5: Cairns KM

Contributions: Manuscript text completed entirely by KMC.

Table of Contents

Abstract I

Acknowledgements III

Academic Acknowledgements VI

Notes on Thesis Structure IX

Author Contributions X

Table of contents XII

List of figures XXI

List of tables XXIII

Overview of thesis chapters XXIV

XII

Chapter 1: Introduction – The identification of dingoes in a background of hybrids

Abstract 2

1.1 Introduction 3

1.2 What is a dingo? 5

1.3 Dingo-European dog hybrids 8

1.4 Why conserve the dingo? 9

1.5 Where are dingoes found? 13

1.6 Genetic and geographic variation within the dingo 15

1.7 Genetics of coat colour and hair texture in the dingo and domestic

dog 16

1.7.1 Agouti 18

1.7.2 MC1R 21

1.7.3 CBD103 23

1.7.1 Coat texture genes 24

1.8 SNP genotyping 25

1.9 Conclusion 31

XIII

Chapter 2: New insights on the history of canids in Oceania based on

mitochondrial and nuclear data 33

Abstract 34

2.1 Introduction 35

2.2 Materials and methods 42

2.2.1 Canid Sampling 42

2.2.1.1 Mitochondrial DNA 42

2.2.1.2 Nuclear DNA 43

2.2.2 Genetic Investigations 43

2.2.2.1 MtDNA genome analysis 43

2.2.2.2 Nuclear gene analysis 44

2.2.3 Phylogeny 45

2.2.3.1 Whole mtDNA phylogenetic analysis 45

2.2.3.2 Nuclear phylogenetic analysis 46

2.2.3.3 Topology testing 47

2.2.4 Statistical analyses 48

2.2.4.1 MtDNA 48

2.2.1.2 Nuclear 49

2.2.5 Estimating divergence time and substitution rate of mtDNA 50

2.3 Results 51

2.3.1 Genetic analyses 51

2.3.1.1 MtDNA genome analysis 51

2.3.1.2 Nuclear gene analysis 51

2.3.2 Phylogeny 52

XIV

2.3.2.1 Whole mtDNA phylogenetic analysis 52

2.3.2.2 Nuclear phylogenetic analysis 55

2.3.2.3 Topology testing 57

2.3.3 Statistical analyses 57

2.3.3.1 MtDNA 57

2.3.3.2 Nuclear 64

2.3.4 Estimating divergence time & substitution rate of mtDNA 65

2.4 Discussion 68

2.4.1 Conclusions 75

Chapter 3: Biogeography of the native dog using uniparental markers: differences

in maternal and paternal biogeography 76

Abstract 77

3.1 Introduction 78

3.2 Materials and methods 84

3.2.1 Canid sampling 84

3.2.2 Mitochondrial gene analysis 86

3.2.3 Y chromosome gene analysis 88

3.2.4 Biogeographic analyses 89

3.2.5 Neutrality tests 91

3.3 Results 91

3.3.1 Biogeographic analyses 91

3.3.2 Neutrality tests 106

3.4 Discussion 110

3.4.1 Conclusions 119

XVI

Chapter 4: Elucidating biogeographical patterns in native canids using whole

genome genotyping technologies 120

Abstract 121

4.1 Introduction 122

4.2 Materials and methods 128

4.2.1 Canid sampling 128

4.2.2 Illumina HD Canine genotyping 129

4.2.3 Inbreeding and homozygosity 129

4.2.4 Clustering analysis 130

4.2.5 Principal components analysis 130

4.2.6 Genetic distances 131

4.2.7 Phylogenetic analyses 131

4.3 Results 133

4.3.1 Illumina HD Canine genotyping 133

4.3.2 Inbreeding and homozygosity 134

4.3.3 Clustering analysis 134

4.3.4 Principal components analysis 140

4.3.5 Genetic distances 140

4.3.6 Phylogenetic analyses 145

4.4 Discussion 150

4.4.1 Conclusions 157

XVII

Chapter 5: General discussion 159

5.1 Introduction 160

5.2 Biogeography 160

5.3 Introduction route, origin and arrival 163

5.4 The mixed identity of Fraser Island dingoes 167

5.5 Hybridisation – evidence of historical paternal introgression 169

5.6 Future of the dingo – conservation and sustainable management 171

5.7 Future directions 174

5.8 Conclusions 175

XVIII

References 176

Chapter 1 – 5 references 177

XIX

Appendix 193

Appendix 1 List of co-authored publications 194

Appendix 2 Canid sample data including geographical sampling co-

ordinates, sample ID, sex and summary of genetic

population data 195

Chapter 2 supplementary material 201

Appendix 3-1 PCR conditions and primers for Whole Mitochondrial

Genome amplification and sequencing of the dingo and

NGSD 201

Appendix 3-2 PCR conditions and primers for Nuclear gene amplification

and sequencing of the dingo and NGSD 205

List of figures

Figure 1-1 Photograph of a dingo from Taronga Zoological Park (Sydney) 7

Figure 1-2 Aboriginal rock paintings depicting Creation Dog 12

Figure 1-3 Map of Australia depicting dingo distribution 14

Figure 1-4 SNP Distribution 30

Figure 2-1 Bayesian analysis of 29 canine mtDNA genomes 54

Figure 2-2 Bayesian analysis of 31 canine concatenated nuclear haplotypes 56

Figure 2-3 Sliding Window Analysis of Dingo mitochondrial genomes 59

Figure 3-1 Map depicting geographic sampling of dingoes across Australia 85

Figure 3-2 Median spanning network based on mitochondrial control region data from

450 dingoes and 23 NGSD 99

Figure 3-3 Median spanning network based upon the mitochondrial diagnostic region

(1706 bp) in 124 dingoes and 5 NGSD. 100

Figure 3-4 Biogeographical map of 120 dingoes and their mitochondrial lineage

designation 101

Figure 3-5 Bayesian analysis of 124 dingo and 5 NGSD mtDNA diagnostic region

(1706 bp) sequences 103

Figure 3-6 Median spanning network based Y chromosome SNP and STR haplotypes

for 79 dingoes and 2 NGSD 104

Figure 3-7 Median spanning network based Y chromosome SNP and STR haplotypes

for 173 dingoes and 20 NGSD 105

Figure 3-8 Biogeographical map of 169 dingoes and their Y chromosome haplogroup

designation 107

XXI

Figure 3-9 Median spanning network based Y chromosome SNP and STR haplotypes

for 173 dingoes, 20 NGSD and 79 South East Asian dogs 108

Figure 4-1 Cross-validation errors for each K-value averaged across ten independent

runs in Admixture Software v1.23 136

Figure 4-2 Maximum likelihood population clustering analysis on 23 dingoes and 5

NGSD. 137

Figure 4-3 Geographical map depicting sampling location of each sample and

majority population cluster identity 139

Figure 4-4-1 Principal Components Analysis based upon filtered whole genome SNP

genotypes for 23 dingoes and 5 NGSD in 3-dimensions 141

Figure 4-4-2 Principal Components Analysis based upon filtered whole genome SNP

genotypes for 23 dingoes and 5 NGSD 142

Figure 4-5 Classical multidimensional scaling analysis with three factors based upon

filtered whole genome SNP genotypes for 23 dingoes and 5 NGSD 143

Figure 4-6 UPGMA tree depicting genetic distances (1-IBS) between 23 dingoes and

5 NGSD 144

Figure 4-7 Maximum Likelihood tree based upon 4913 ancestry informative markers

in 23 dingoes and 5 NGSD 147

Figure 4-8 Maximum likelihood tree constructed based upon 6,288 informative SNPs

in 23 dingoes, 5 NGSD and 12 wolves. 148

Figure 5-1 Map depicting hypotheses of the origin and introduction route of the

dingo. 166

XXII

List of tables

Table 1-1 Coat color patterns and genetics in the dingo. 19

Table 1-2 Differences between 454 dingo sequence and Boxer reference sequence on

chromosome 1 (5% of the dog genome) 29

Table 2-1 Descriptive measures of Nucleotide diversity for mtDNA and nuclear gene

regions 53

Table 2-2 Topology testing to investigate the monophyly of dingoes using Bayes

factors 58

Table 2-3 Neutrality test results for mitochondrial and nuclear gene regions 61

Table 2-4 FST values for MtDNA and nuclear gene regions 62

Table 2-5 AMOVA results for mtDNA and nuclear loci 63

Table 2-6 Estimates of divergence times of dingo lineages 67

Table 3-1 PCR amplification primers for mitochondrial DNA amplification and

sequencing in the dingo and NGSD 87

Table 3-2 Mitochondrial and Y chromosome haplotype data for 124 dingoes and 5

NGSD 93

Table 3-3 Nucleotide variation and Neutrality statistics on mitochondrial DNA from

124 dingoes. 109

Table 4-1 Inbreeding (FIS) and X chromosome homozygosity (F) for 23 dingoes and

5 NGSD calculated from Illumina SNP data 135

Table 4-2 FST statistics calculated between the dingo and NGSD populations 138

Table 4-3 Topology testing between unconstrained and constrained maximum

likelihood analyses 149

XXIII

Overview of thesis chapters

It is believed by conservation groups that there are different ecotypes of dingo, alpine, desert and/or tropical, however little scientific research has explored this question. The primary goal of this thesis is to investigate the presence of genetic subdivision and/or variation within the Australian dingo using a range of genetic markers. This thesis also aims to investigate the origins, evolutionary history, patterns of dispersal and modern biogeography of the dingo. These questions have important conservation and management implications. The null hypothesis investigated in this thesis is that dingoes are a homogeneous population with no geographical subdivision and immigrated to

Australia once from South East Asia. An alternate hypothesis is that there is geographic subdivision in the dingo, a result of adaption to the Australian environment. A second alternate hypothesis is that geographic subdivision is present in the dingo and is a result of multiple introductions into Australia. These alternate hypotheses may be investigated through estimation of neutrality statistics and exploring the phylogenetic relationships between dingoes and NGSD or Asian dogs across multiple genetic markers.

This thesis contains five chapters. Chapters 1 to 4 are written as separate manuscripts for publication. Chapter 1 is a literature review. Chapters 2 to 4 are research chapters, with abstract, introduction, methods, results and discussion. At the time of submission, these chapters were manuscripts in preparation (Chapters 2, 3 and 4) or published

(Chapters 1). Chapter 5 is a general discussion chapter that synopsises the main conclusions of the thesis, with conservation and management implications, and proposes future research directions to develop these findings.

XXIV

Chapter 1 presents a review of the current knowledge concerning the dingo. The review outlines the taxonomy and distinguishing features of the dingo followed by a review of the primary current conservation concern: hybridisation with domestic dogs. The focus then shifts to the ecological, economic and cultural importance of the dingo followed by a discussion of the distribution of the dingo in Australia. Current patterns of genetic and geographic variation are explored followed by a discussion of the genetics of coat colour in dingoes and dogs. The final review portion of the chapter focuses upon the utility of single nucleotide polymorphism (SNP) technologies to exploring genetic variation and hybridisation in the dingo.

Chapter 2 explores the presence of geographic subdivision in the dingo using whole mitochondrial genomes and thirteen nuclear loci. It identified evidence of two distinct populations of dingo: southeastern (SE) and northwestern (NW). Additionally, evidence that the dingo lineages likely diverged outside Australia is presented, indicating multiple introductions. Mitochondrial divergence time estimates suggested dingoes may have arrived in Australia earlier than previously reported.

Chapter 3 focuses upon a broader survey of genetic variation in dingoes utilising mitochondrial and Y chromosome markers. This study also observed two distinct lineages of dingoes. However, mixing between the lineages was observed more commonly at the paternal markers suggesting differences in female versus male dispersal rates. Additionally evidence of paternal introgression from domestic dogs was observed in the SE lineage of dingo.

XXV

In Chapter 4, the thesis explores the presence of geographic subdivision in dingoes using whole genome SNP data. Three distinct dingo populations were observed: SE,

Fraser Island (FI) and NW. The FI dingo population whilst distinct appears to be the result of an initial SE introduction followed by introgression by NW paternal lineages.

It is possible that human movements between the Australian mainland and Fraser Island may have facilitated this paternal gene flow. The FI population appears to be compromised by inbreeding. NGSD were observed to be inbred and divergent from dingoes, not unexpected given the small foundation size of the captive NGSD population. A set of diagnostic SNP markers has been identified for future genetic surveys of dingoes to explore biogeography.

The final chapter of this thesis, Chapter 5, discusses the conservation and management implications in the light of the observed population structure. Particularly evident is the need to adjust current strategies to maintain and protect the existing population structure. Male biased introgression from domestic dogs appears to be a particular issue for the SE dingo population, suggesting the need to inhibit further hybridisation in southeastern Australia. The FI dingo population appears to be inbred; indicating the need for more in depth genetic research concerning this population, so that suitable management strategies can be developed. NW dingoes appear to be widespread across mainland Australia.

XXVI

Chapter 1

(Introduction)

The identification of dingoes in a background of hybrids

1 Abstract

The dingo is an ancient dog that has been isolated on the Australian continent for approximately 5,000 years. Hybridisation with European domestic dogs is threatening the existence of dingoes and has the potential to compromise conservation strategies.

Physical characteristics of the dingo, such as size, colour and markings, have been used in an attempt to differentiate dingoes from hybrids. However, morphological characteristics can be unreliable. This has the potential to impede conservation efforts, as hybrids cannot be selectively excluded from breeding programs based upon morphology. We suggest that genetic investigation of coat colour and hair texture loci may help resolve uncertainties concerning possible colour variations in the Australian native dog. Microsatellite data are currently used to differentiate dingoes from hybrids but genome coverage is incomplete. Single nucleotide polymorphisms (SNPs) are likely to be an effective tool to quantify geographic variation as well as the extent of dingo- domestic dog hybridisation, as a large number of SNPs can be typed simply and economically for each sample. Commercial canine SNP arrays contain many markers that are informative for differentiating ancient dogs like the dingo from modern domestic dog breeds, but these SNPs are mainly chosen based on polymorphism in modern domestic dogs. Future studies may aim to identify 1000 SNPs from partial sequence of the dingo genome, which will be highly informative for determining the occurrence and extent of dog gene introgression in a dingo. Such markers could also be used to follow changes in the dingo population over time and identify regions under selection. In combination, the information is expected to add contextual data to the conservation efforts to save the native dog of Australia.

2 The greatest problem in trying to write about the dingo is that one has no proof of the animal's

identification, ancestry, affinity, place of origin, or precise time of arrival in Australia; even

questions of whether it was on arrival a partly domesticated animal, subsequently becoming

feral, or whether it is truly a wild dog, is disputed by many writers, and on almost completely

polemical grounds (Macintosh 1975).

1.1 Introduction

The taxonomy of the dingo, or Australian native dog, is under debate, with some referring to it as Canis lupus dingo, a subspecies of wolf, while most recently it has been classified as Canis lupus familiaris dingo, a breed of domestic dog (Fleming et al.

2001; Jones 1925; Macintosh 1975; Walters 1995). It is considered to be an ancient breed, indicative of the ancestral undifferentiated domestic dog (Corbett 1995) and, as such, provide a window into the evolution of the domestic canine. They are closely related to other ancient dog breeds of Asian origin, such as the New Guinea Singing

Dog, Akita, Shar-pei, Chow-Chow and Basenji (Koler-Matznick et al. 2004; Puja et al.

2005; vonHoldt et al. 2010; Walters 1995).

There is considerable debate about the origin of dogs, including dingoes. Mitochondrial

DNA (mtDNA) data suggests a single point of origin (Pang et al. 2009; Savolainen et al. 2002). Savolainen et al. (2002) and Pang et al. (2009) place this origin in southern

China because dogs in this region have the highest mtDNA variability. VonHoldt et al.

(2010) suggest instead that this origin was in the Middle East due to the higher level of nuclear DNA single nucleotide polymorphism (SNP) haplotype sharing between domestic dogs and wolves from this region. However, the authors found that ancient

3 Asian breeds, including the dingo, are highly divergent from modern European dog breeds and show genetic patterns consistent with admixture or evolution from Asian wolves. In combination, the data suggests that dingoes evolved in Asia either from

Asian wolves or as a result of hybridisation with Asian wolves, and are genetically distinct from modern European dog breeds (Pang et al. 2009; Savolainen et al. 2002,

2004; vonHoldt et al. 2010).

The Australian native dog has been present on the continent for 3,000-5,000 years according to both fossil (Gollan 1984; Jones 1925; Macintosh 1975) and DNA evidence

(Savolainen et al. 2004; vonHoldt et al. 2010). MtDNA data suggest that a small number of animals colonised Australia (Savolainen et al. 2004). Y chromosome SNPs support the hypothesis that there were a low number of introductions (Peter Savolainen, pers. comm.). It is plausible that seafaring traders whom might have kept ancestral dingoes as a source of food or as guard dogs during long voyages, brought dingoes to

Australia (Corbett 1995; Fleming et al. 2001). These seafarers likely introduced the ancestral dingoes to the northern coastal regions from which they spread to the southern regions of the continent (Corbett 1995; Fleming et al. 2001).

Prior to European colonisation, dingoes were present as tamed camp canines as well as free-living animals (Fleming et al. 2001; Walters 1995). Aboriginal Australians had different words for the two. In some languages the camp dingoes are “tingo”,

“warrigal”, or “maliki”. The free-living counterparts are referred to as “wantiburri” or

“warrukadli” (Fleming et al. 2001; Jones 1925; Macintosh 1975). It is clear that the

Australian native dog played an important role in Aboriginal culture and its

4 spiritualism, with many dreamtime stories and rock art galleries being focused on the animal (Macintosh 1975; Smith & Litchfield 2009). After the First Fleet colonised

Australia in 1788, many Aboriginals abandoned their camp canines either as a result of having to flee in the face of colonial violence or because the European domestic dogs were perceived to be better companions (Breckwoldt 1988; Smith & Litchfield 2009;

Walters 1995).

Current national and state laws in Australia are ambiguous, and perhaps even contradictory, as to whether dingoes should be conserved or shot (Fleming et al. 2001).

In rural Australia, wild dogs are vilified as livestock killers. Indeed, active culling and baiting programs are extensive in sheep and cattle country (Fleming et al. 2001). In contrast, in cities native dogs are seen as an Australian icon, with public outrage over their mistreatment on certain islands (Anonymous 2010) and disbelief that a dingo could attack a human baby (Brown 2010). Nation-wide conservation groups are fighting to keep the dingo alive by educating the public and running captive breeding programs

(Walters 1995).

1.2 What is a dingo?

The general appearance of the Australian native dog is that of a medium built, elegant and active dog of

great nimbleness and agility. It has a short straight back with distinct waist and cut-up, and gives the

impression of being "high on the leg". Aristocratic and graceful, the breed clearly displays its purity and

nobility. Excerpt from ANKC breed standard (Walters 1995).

5 The dingo is a primitive medium-sized dog (Figure 1-1), which weighs between 10-20 kg when fully grown (Jones 1925; Kaleski 1933; Walters 1995). All dingoes have a short straight coat, with hairs averaging 30-50 mm in length. Ears are erect and hooded while the tail can be either curled over the back or hanging down. The tail is generally described as bushy. Dingoes have almond-shaped eyes, which combined with their erect ears and skull structure give them a distinctly fox-like appearance. Their eyes are generally a deep brown to hazel colour and their nose, eye rim and paw pad pigment is generally black (Jones 1925; Kaleski 1933; Walters 1995).

The dingo has several characteristics that are indicative of their wild and ancient origin.

It has large carnassial and canine teeth, suggestive of a wild predator that hunts for its food (Jones 1925; Macintosh 1975). Most domestic dogs have smaller carnassial and canine teeth as a result of domestication (Jones 1925). Dingoes have a skull shape that is relatively primitive and undifferentiated, and closer in shape to many wolves than most modern dog breeds (Corbett 1995, 2001; Macintosh 1975; Newsome et al. 1980).

Animals of both sexes have a single annual mating season extending from late autumn to mid-winter, with pups being born in late winter (Catling et al. 1992; Corbett 1995;

Corbett & Newsome 1975; Walters 1995). A single annual breeding season is a hallmark of most wild animals and, more specifically, of most (if not all) ancient dog breeds such as the Basenji (Johannes 2003). Modern European domestic dogs, on the other hand, typically come into oestrous twice a year, potentially because they have evolved with a more abundant supply of food (Catling et al. 1992).

Figure 1-1 Photograph of a dingo from Taronga Zoological Park (Sydney), showing typical dingo morphology (see above). Photo reprinted with the permission of the JL

Shellshear Museum Archives (University of Sydney).

7 Dingoes can bark, but this is not common (Corbett & Newsome 1975; Walters 1995).

Instead, dingoes yodel and howl in a similar way to other ancient breeds such as the

Basenji and New Guinea Singing Dog (Corbett & Newsome 1975; O'Neill 2002;

Walters 1995). This yodel or howl has been interpreted to be a more complex communication ‘language’ than the bark of domestic dogs (Corbett & Newsome 1975;

Walters 1995). Dingoes have demonstrated high problem-solving abilities by completing detour tasks in shorter time spans and with fewer errors than domestic dogs of varying breeds (Smith & Litchfield 2010).

1.3 Dingo-European domestic dog hybrids

Dingo-European domestic dog hybridisation has been ongoing since 1788 when

European colonialists transported their companion canines to Australia (Breckwoldt

1988; Walters 1995). Despite this evolutionarily short period of sympatry, hybridisation has been extensive (Corbett 1995, 2001; Elledge et al. 2006, 2008; Jones 2009;

Newsome & Corbett 1985; Wilton 2001; Wilton et al. 1999). One plausible explanation for the extent of the hybridisation is that culling and baiting of dingoes breaks up pack structures, making them more likely to associate with domestic dogs (Wallach et al.

2009). This hypothesis is supported by DNA data, which suggest that a majority of hybrids result from the mating of female dingoes and male domestic dogs. In a stable dingo pack, only the alpha female and male mate, and it is unlikely that a female dingo would mate with a wandering domestic dog (Savolainen et al. 2004). An alternate hypothesis is that population control has resulted in population size declines and this

8 has facilitated genetic bottlenecking, allowing domestic dog alleles to become fixed in dingo populations.

In the 1990s a genetic testing method using microsatellite markers was developed to detect hybrids (Wilton 2001; Wilton et al. 1999). A set of 23 markers was chosen based on the ability to discriminate between dingoes and domestic dogs. This method has been used successfully to identify dingo-domestic dog hybridisation events that have occurred within the last four generations. However, this set of 23 markers does not give complete coverage of all 39 canine chromosomes (Elledge et al. 2008; Wilton 2001).

This has the potential to be a problem if the rates of introgression from domestic dog into dingo are not random (Elledge et al. 2006, 2008; Wilton 2001; Wilton et al. 1999).

Livestock predation causes direct economic losses to farmers, and also leads to indirect costs associated with canine control (Catling et al. 1992; Fleming et al. 2001; Newsome

& Corbett 1982). As an example, in 1988 the cost of canine predation to the sheep industry in eastern New South Wales was estimated at over $4 million (Fleming et al.

2001). It is popularly believed that dingo-domestic dog hybrids are ecologically and economically more deleterious to farmers than dingoes, as they may be larger in size and more aggressive than dingoes; however ecological research needs to test such conjecture.

1.4 Why conserve the dingo?

There are compelling arguments that the dingo should be conserved for scientific,

9 ecological, cultural and economic reasons. Biologically, dingoes provide insight into the origins, behaviour and appearance of early dogs. Native dogs can provide a genetic tool for studying the genes affected by domestication (vonHoldt et al. 2010). As an undifferentiated and ancient dog, the dingo provides a good comparison for investigating the genes that have been selected by humans to produce the modern breeds we know today, e.g. the brachycephalic Pug, diminutive Yorkshire terrier or immense

Great Dane. Such investigation may facilitate the understanding of genes that affect skull morphology, dentition, coat type, body size, etc. and thereby identify diseases and traits associated with these genes (Chase et al. 2009; Shearin & Ostrander 2010). The non-domesticated behaviour and temperament of the dingo may also prove an important tool, as comparison with modern dog breeds may pinpoint genes underlying the transition from wild animal to perfect pet in terms of behaviour and temperament

(Chase et al. 2009).

Ecologically, dingoes fill the niche of top terrestrial predator within Australia. Without dingoes, severe ecological repercussions can occur, as there is no top-down control of populations of large herbivores and invasive mesopredators. Research on the abundance of small marsupials in the presence of dingoes has shown a clear beneficial effect

(Fleming et al. 2001; Glen & Dickman 2005; Glen et al. 2007; Johnson et al. 2007;

Johnson & Wroe 2003; Letnic & Koch 2010; Letnic et al. 2009; May & Norton 1996).

Dingoes limit the numbers of large herbivores such as the Red Kangaroo (Fleming et al.

2001; Letnic & Koch 2010; Letnic et al. 2009; Pople et al. 2000) and indirectly protect many small-medium native species by excluding introduced invasive mesopredators such as the fox (Glen & Dickman 2005; Glen et al. 2007; Johnson et al. 2007; Letnic et

10 al. 2009, 2010; Letnic & Koch 2010). Dingoes also prey upon feral herbivores including pigs and goats (Fleming et al. 2001; Glen et al. 2007).

The Australian native dog plays an important cultural role in indigenous Australian culture. The animal is a common feature in Australian Aboriginal dreamtime stories, which are an important part of the indigenous culture, spiritualism and oral history

(Cairns & Harney 2004; Harney 1951; Smith & Litchfield 2009). One example is a

Cape York dreamtime story of the Giant Devil Dingo who becomes man’s friend and helper (Roughsey 1973). Dingoes are frequent subjects of rock art, paintings and rock engravings found across the continent (Figure 1-2) (Smith & Litchfield 2009).

Economically, the dingo is responsible for significant economic benefit in the form of tourism. International and national tourists flock to zoos and dingo refuges to see the canid first-hand (Fleming et al. 2001; Gray 2010). Dingoes are considered by visitors to be an Australian icon, as unique to the continent as kangaroos, koalas and wombats

(Fleming et al. 2001; Smith & Litchfield 2009).

Figure 1-2 Aboriginal rock paintings depicting Creation Dog, a totemic animal that is strongly connected to the dingo in Aboriginal

culture. According to Wardaman elder Bill Harney “You know Creation Dogs are a dingo.” (Cairns & Harney 2004). Photographs are

reprinted with the permission of publisher, HC Cairns.

1.5 Where are dingoes found?

Historically, the Australian native dog could be found all across mainland Australia, although they were particularly bountiful in the coastal regions and areas with natural water sources. Native dogs have never inhabited Tasmania and this is likely because they did not arrive in Australia until after the land bridge between the mainland and

Tasmania was closed 12,000 years ago (Breckwoldt 1988; Corbett 1995; Fleming et al.

2001).

In the early 1860s the dingo and dingo-domestic dog hybrid populations rapidly increased due the spread of rabbits throughout Australia (Jones 1925) and the introduction of man-made watering holes (Fleming et al. 2001). This rapid increase in dingo populations prompted one of the largest undertakings in Australian history: the building of the ‘dingo fence’ (Figure 1-3). The dingo fence is a 5,614 km pest-exclusion fence, initially built during the 1880s as a barrier to rabbits. In 1912-1918 the various rabbit fences were joined to create a dingo proof barrier. The dingo fence was fully completed in the 1950s preventing dingoes from entering the fertile southeastern region of Australia thereby protecting the sheep industry in New South Wales, Victoria and

Queensland. Today, dingoes are relatively common in most of central, northern and western Australia (Breckwoldt 1988; Corbett 1995; Fleming et al. 2001) and are present on some Australian islands, e.g. Fraser Island (Fleming et al. 2001; Jones 2009) (Figure

1-3). Small fragmented dingo populations do exist south of the dingo fence; however, landholders heavily control these to prevent extensive dingo predation (Breckwoldt

1988; Fleming et al. 2001).

Figure 1-3 Map of Australia depicting dingo distribution and location of the dingo fence, shown by a thick black line. Figure adapted from Fleming et al. (2001).

14 1.6 Genetic and geographic variation within the dingo

The outward appearance varies considerably from the sinewy, single-coated, rippling muscled dog of the

tropical far north, to the fox-like, or coyote-like dog of the colder southern mountain regions, which has a

thick double coat, or the smaller and finer dog of the arid regions. The coat is seasonal. Excerpt from

ANKC breed standard (Walters 1995).

The breed description in the Australian Kennel Club breed standard (quoted above) implies that there is morphological variation in the dingo. Colloquially, three dingo varieties are thought to occur: tropical, desert, and alpine. These varieties are heuristically distinguished from each other based on size, stature, coat colour and texture of coat. It is currently not known whether this morphological variation has any genetic basis or whether it is caused by adaptation to local conditions (Bueler 1974).

Since the 1860s the question of whether coat colour and texture differ between dingo varieties has been debated (Corbett 1995; Walters 1995). It has been suggested that alpine dingoes are more likely to be white, while tropical dingoes are more likely to be black, or black and tan. Very few historical accounts of dingo colouration are available from indigenous sources. Accordingly, for colourations other than ginger, it is unknown whether they were described after very early hybridisation events or were present before

European arrival (Bueler 1974; Gould 1863; Jones 1925; Macintosh 1975).

Today, conservation groups and scientists support the existence of black and black-and- tan dingoes (Breckwoldt 1988; Bueler 1974; Fleming et al. 2001; Jones 1925;

Macintosh 1975; Newsome & Corbett 1985; Walters 1995), but others say the native

15 dog can only be ginger or red (Corbett 2001; Kaleski 1933). The presence of black and/or black-and-tan colouration in closely related breeds such as the Thai pariah dog,

New Guinea Singing Dog, Shiba Inu and the Basenji suggests that the presence of black colouration may be ancestral in dingoes.

A method for distinguishing dingoes from hybrids is required by conservation managers to selectively choose animals for captive breeding programs. To date, scientists and conservationists alike have used physical diagnostic characters such as colouration, skull morphology and breeding habits (Corbett 1995, 2001; Elledge et al. 2008; Jones

1925; Macintosh 1975; Newsome et al. 1980) to assess the extent of hybridisation.

However, these methods are fraught with complications and errors, not least of which is that skull morphology measurements must be completed on dead animals (Elledge et al.

2006, 2008; Jones 2009).

1.7 Genetics of coat colour and hair texture in the dingo and domestic dog

One scientific approach to move the debate forward is to understand the colour and hair texture genetics of dingoes. This is important for conservation purposes for three main reasons: (1) to inform human perceptions, (2) to investigate hypothesised geographical and genetic variation, and (3) to examine introgression. Australians today have a distinct perception of the physical characteristics of the native dog. These perceptions underpin modern conservation and control practices. Unfortunately, the scientific rigor of these discussions is impeded by the lack of empirical data.

16 A major objective of our research program is to critically inform debates concerning the genetic variation within the dingo. Investigation of genetic variation in the genes involved in coat colour and texture and the critical evaluation of SNPs in the chromosomal regions surrounding colour genes is expected to enable research concerning the extent, time span and origin of dog introgression. Notably, coat colour genetics has been used to investigate introgression in gray wolves, with melanism

(black colour) being identified as a result of domestic dog introgression (Anderson et al.

2009; Schmutz et al. 2007). Additionally, genetic testing of the colour genotypes along with other genetic markers, such as SNPs, is likely to provide estimates of the extent of introgression if recessive or cryptic dog-only colour genes are present.

In this section, we consider the genes likely to be involved in colouration of the dingo, followed by a description of hair texture genes. There are nine different genes that are known to control colouration in the dog (Anderson et al. 2009; Berryere et al. 2005;

Candille et al. 2007; Dreger & Schmutz 2010; Everts et al. 2000; Hong et al. 2009;

Kerns et al. 2003, 2004, 2007; Little 1957; Newton et al. 2000; Philipp et al. 2005;

Rothschild et al. 2006; Schmutz & Berryere 2007; Schmutz et al. 2003, 2007, 2009).

However, the genes that are most likely to affect coat colour in dingoes are Agouti (A locus), MC1R (E locus) and CBD103 (K locus) (Table 1-1). These three genes interact to control the pattern and type of pigmentation present in a dog’s coat and have been found to affect coat colour in mice (Bultman et al. 1994; Millar et al. 1995; Siracusa

1994; Vrieling et al. 1994), pigs (Drögemüller et al. 2006), sheep (Norris & Whan

2008) and cattle (Miao et al. 2010).

17 1.7.1 Agouti

Agouti or Agouti signal peptide (ASIP) controls the distribution of black (eumelanin) and red (pheomelanin) pigment in a dog’s coat. It has been mapped to chromosome 24 and spans nearly 35,000 base pairs (bp). Currently, it is believed that there are four

Agouti exons, three coding and one untranslated region. It has been speculated that several unidentified untranslated exons exist and may play a role in the laying down of pigment during the hair growth cycle or cause ventral-specific expression (Schmutz &

Berryere 2007).

The four Agouti alleles as described by Little (1957), and corroborated by more recent genetic evidence, are ay, aw, at and a in order of dominance (Berryere et al. 2005; Kerns et al. 2004; Schmutz & Berryere 2007). Some dog breeders believe a fifth allele (as) exists; however, there is no genetic evidence of this allele to date (Schmutz & Berryere

2007). In dingoes it is likely that the ay and at alleles exist in current wild and captive populations. It is plausible that the aw and a alleles also exist, but these alleles are likely to be rare in both wild and captive populations.

Table 1-1 Coat colour patterns and genetics in the dingo. Descriptions of phenotypes and likely genotypes of possible dingo colour varieties Colour Description Alleles Ginger A dingo with a red coat varying in (1) ayay but may be EE, shade from deep rust to pale cream. Ee or ee and must be No interspersed black hairs present kyky along spine. Often white colouration (2) KBKB and ee with on muzzle, neck, chest, feet, tail tip, any Agouti genotype etc. Sable A dingo with a red coat varying in (1) ayay shade from deep rust to pale cream. (2) awaw Black hairs present along spine or (3) ayat or ayaw across entire body. Often white Must be EE or Ee and colouration on muzzle, neck, chest, kyky feet, tail tip, etc. Black-and- A black dingo with tan points, brow (1) atat with EE or Ee tan pips, muzzle, chest, belly, feet and and kyky legs. Tan may vary from deep tan to a cream. White A dingo with a solid white coat. May (1) aa with ee and kyky have lighter nose, paw pad and eye (2) ayay with ee and rim pigmentation. kyky (3) KBKB with ee and any Agouti genotype and an unknown dilution modifier Solid Black A dingo with a solid black coat. (1) aa with EE or Ee May have white markings on and kyky muzzle, neck, chest, feet, tail tip, etc. (2) KBKB orKBkbr or KBky with EE or Ee and any Agouti genotype Brindle A dingo with an alternating black (1) ayay or aya* with and red striped pattern. kbrkbr or kbrky and EE or Ee * Any allele of lesser dominance

19 The ay allele is sable where an animal has red/yellow hair that may or may not have black tipping. The ay allele contains two amino acid changes compared to the wild type,

A82S (246G>T), from an Alanine to a Serine and R83H (250G>A), from an Arginine to a Histidine (Berryere et al. 2005; Kerns et al. 2004). Homozygote ayay likely causes the common ginger dingo colouration pattern. Ginger dingoes can have black or black- tipped guard hairs along their spine, often called sabling (Walters 1995); however, it is unclear whether this is due to the actions of ay, aw or at alleles or a modifier gene.

The aw allele is wild type and it describes the colour pattern of yellow and black banded hairs. This allele is often termed wolf-sable or agouti, and is the natural colouration of wild gray wolves (Anderson et al. 2009; Berryere et al. 2005; Kerns et al. 2004;

Schmutz & Berryere 2007; Schmutz et al. 2007). It is plausible that this allele causes sable colouration in dingoes. It is possible that aw in dingoes either results in dingoes that are a full sable colour (similar to wolf gray) or are ginger with black-tipped hairs along the spine. It is equally possible that sabling, black or black-tipped guard hairs in dingoes is due to the ay and at alleles and that aw is not present in dingoes.

The at allele describes the black-and-tan point colour pattern. In mice, the aw and at alleles differ only in the promoter region (Candille et al. 2004; Millar et al. 1995;

Siracusa 1994; Vrieling et al. 1994) and it is theorised that this mechanism controls the agouti versus black-and-tan colour patterns in dogs (Schmutz & Berryere 2007). In dingoes, the black-and-tan colour pattern is likely caused by a homozygote atat genotype. Heterozygote expression of at with ay may be the cause of a ginger colouration with black hairs interspersed along the spine, called sabling as noted above,

20 as is the case in the Australian Cattle Dog, a breed with dingo heritage (Hewson-Freund

2003).

The recessive black allele a causes solid black coat colour. The a allele is caused by a substitution mutation (288C>T) resulting in an amino acid change, R96C from Arginine to Cysteine (Berryere et al. 2005; Kerns et al. 2004). It is probable that a is the allele responsible for solid black colouration in dingoes; however, as individuals with this colouration are extremely rare, such investigation may be difficult without the use of museum or historical specimens. The rarity of black dingoes could be due to selection against black dingoes combined with the recessive nature of the allele. Dingoes with an aa genotype may have white markings and consequently may be confused with black- and-tan dingoes.

1.7.2 MC1R

Melanocortin 1 receptor (MC1R), more commonly known as the extension (E) locus, controls whether an animal is capable of producing both red and black pigment or red pigment only (Newton et al. 2000; Schmutz & Berryere 2007). MC1R mutations are epistatic to Agouti mutations. MC1R has been mapped to chromosome 5 and is a single exon gene approximately 950 bp in length (Everts et al. 2000; Schmutz & Berryere

2007).

There are four known alleles at MC1R. In order of decreasing dominance the alleles are

E, EM and e, while EG is dominant to E and e but recessive to EM (Dreger & Schmutz

21 2010; Everts et al. 2000; Hong et al. 2009; Kerns et al. 2003; Schmutz & Berryere

2007). It is postulated that MC1R E and e alleles exist in wild and captive dingo populations, but not EM or EG.

E is the wild type extension allele, and dogs with this allele can produce both red and black pigment (Everts et al. 2000; Newton et al. 2000; Schmutz & Berryere 2007). In terms of phenotypic variation, an EE or Ee genotype dingo would appear as the Agouti gene determined, i.e. both EE atat and Ee atat dogs would be black-and-tan.

The recessive allele of the extension locus is e, which prevents the dog from producing black pigment; a homozygote has a red or white coloured coat. The e allele is caused by a loss of function mutation R306ter (914C>T), which causes a premature stop codon to replace the usual Arginine (Everts et al. 2000; Newton et al. 2000; Schmutz & Berryere

2007). If a dingo has an ee genotype at MC1R, then black pigment is no longer produced. It is hypothesised that ee may affect a dingo’s colouration in two different ways: (1) ee genotype causes the dingo to be white regardless of Agouti genotype; or

(2) ee genotype causes atat and aa dogs to be white but does not alter the colour of ayay dogs beyond removing the presence of black-tipped or black guard hairs along the spine. It is also possible that white dogs have an ee genotype but must also have a dilution or modifier gene that dilutes red/yellow pigment (Hong et al. 2009; Schmutz &

Berryere 2007).

22 1.7.3 CBD103

Beta-defensin 103 (CBD103) is the gene that causes dominant black and brindle colouration in dogs (Candille et al. 2007; Kerns et al. 2003, 2007; Schmutz & Berryere

2007). CBD103, also known as the K locus, is located on chromosome 16 and has three known alleles, KB, kbr and ky, in order of decreasing dominance. It appears most likely that dingoes are fixed at the K locus for kyky.

It is, however, possible that KB is ancestral in dingoes and causes solid black colouration. It is also plausible that ginger or white dingoes are the result of an interaction between KBKB and ee genotypes. The KB allele causes a dog to have a solid black coat colour in the presence of an E or EM allele at MC1R regardless of Agouti genotype. The associated mutation for the KB allele is a 3 bp deletion, G23, within the

CBD103 coding region, causing the deletion of a Glycine amino acid (Anderson et al.

2009; Candille et al. 2007).

The second CBD103 allele is kbr, which is caused by a segmental duplication event that leads to variable expression of black and red pigment, resulting in a brindle striped appearance (Candille 2009). The kbr allele is expressed only in areas where Agouti is expressed, and only in the presence of an E or EM MC1R allele. That is, in an atat dog only the ventral (tan) areas are brindle, but in an ayay dog the entire body is brindle

(Candille 2009; Schmutz & Berryere 2007). Early reports of brindle dingoes suggest that it is plausible for the kbr allele to be ancestral in origin. The rarity of brindle in dingoes could be explained by ee genotypes that would mask the kbr allele, and

23 extensive culling of exotic coloured dingoes that are generally believed to be hybrids.

Conversely, given the rarity of brindle colouration and lack of pre-European reports of brindle dogs, the kbr allele in dingoes may provide evidence of dog introgression.

The third and final K locus allele is ky, which is the wild type. In a kyky individual, normal expression of Agouti and MC1R occurs (Anderson et al. 2009; Candille et al.

2007; Schmutz & Berryere 2007).

1.7.4 Coat texture genes

Investigation of genes related to hair texture and length may also be helpful in unravelling the question of whether different ‘varieties’ of dingo exist. This may help conservation groups develop captive breeding programs to conserve the entire range of genetic variation in the dingo. It has been postulated by conservation groups that different dingo ‘varieties’ or sub-types differ not only in coat colour but also in coat type; specifically, the length, texture and density of hairs (Walters 1995). Looking at genes controlling hair length and texture combined with morphological microscopy may allow for scientific definition and differentiation of dingoes from separate environmental varieties, if they exist.

Mutations in RSPO2 (which cause wiry hair particularly ‘furnishings’), FGF5 (which causes long and silky coat types) and KRT71 (which causes curly coated dogs) may all prove useful in studying the basis of coat variation in the dingo (Shearin & Ostrander

2010). It is likely that FGF5 and KRT71 would be fixed as wild type as no long, curly haired dingo has been described. If the genes relating to whether a dog has a double coat

24 (guard hairs and undercoat) or single coat (only guard hairs) could be identified they may be most useful in investigating coat type differences between dingoes.

1.8 SNP genotyping

Larger numbers of loci are needed to provide coverage of the dingo genome and accurately determine genetic identity. SNPs can be genotyped in very large numbers

(100,000s) in a single reaction for each sample, at a low cost per SNP. SNPs are common in the dog genome and are available at loci spread across the whole genome, so they can be chosen to give even representation of the genome. Diagnostic loci, where dingoes have one allele and domestic dogs another (or others), would be ideal markers for identifying hybrids and introgression of dog genes into the dingo. A clear advantage of using SNPs is that many can be tested in a single experiment. The major disadvantage is that the upfront costs tend to be high, it is costly to run each sample, and access to higher end computers is required. Certainly, however, testing 100 SNPs would provide a higher level of power to identify hybrids with low levels of dog ancestry compared to the current 23 microsatellite testing method.

Most of the commercially described SNPs have been ascertained by virtue of their polymorphic nature within domestic dogs. Commonly available SNP genotyping platforms for dogs can be sourced from Affymetrix (www.affymetrix.com) and Illumina

(www.illumina.com). The Affymetrix Canine SNP Chip ver. 2 has about 50,000 useable loci (Lindblad-Toh et al. 2005) that can be tested for about $500 per animal. This platform has been used to compare 85 domestic dog breeds, dingoes and wolves

25 (vonHoldt et al. 2010), and the data suggest that the dingo is the most distinct of all of the ancient dog breeds. In dingoes, approximately 60% of alleles at the 100 most informative loci for differentiating wolves and dogs are wolf-like rather than domestic dog-like. The Illumina High Density SNP Array carries 170,000 SNPs. It shares many

SNP markers with the Affymetrix chip, but gives a better and more even coverage of the dog genome. However, only a small proportion of these are likely to be useful for differentiating between dingo and domestic dog origins. Commercially available SNP chips may, however, be useful in identifying geographic population structures in the dingo population.

To investigate dingo-domestic dog introgression, custom SNP chips would need to be utilised. Custom SNP chips can be developed on both Affymetrix and Illumina platforms for very large genotyping projects of thousands of samples. Agilent custom arrays are being developed for SNP genotyping, which will make testing of small numbers of samples practical. Before developing a custom chip, a database of diagnostic SNPs differing between dingoes and domestic dogs needs to be developed.

Diagnostic loci are likely to be relatively common between dingoes and dogs because of the long-term isolation of dingoes on Australia, separate from modern European dogs and the history of population bottlenecks in both lineages (Lindblad-Toh et al. 2005;

Pang et al. 2009; Savolainen et al. 2004). Loci heterozygous in ancestral dogs may have become fixed for alternative alleles in dingoes and domestic dogs. Thus, the well- established isolation between dingoes and domestic dogs can continue to be utilised when defining the Australian native dog. Specifically, two diagnostic loci have already been included among the standard genetic testing set after they were identified during a

26 small sequencing project on random dog genes. One of the diagnostic loci is a haplotype with a two bp (TT) insertion and six SNPs within 100 bps in intron 2 of the PNKD gene on CFA37. The insertion and six SNPs are fixed in dingoes while the alternative haplotype is fixed in modern European domestic dogs, but not all ancient dog breeds or wolves (Britt-Louise Carlsson, pers. comm.).

In an attempt to identify additional dingo- versus domestic dog-informative SNPs, Next

Generation Sequencing has been used to sequence 80 Megabases (Mb), or 3% of the dingo genome in a single 454 run. Alignment with the dog reference genome build v2.1

(Candille 2009) identified 10,000 sites spread across the genome where the 454 sequence differs from the Boxer reference sequence (Table 1-2 and Figure 1-4). The majority (>80%) of these sites were not listed on the SNP database (dbSNP - www.ncbi.nlm.nih.gov/snp) (Figure 1-4), suggesting they may be dingo specific. If, however, just 10% of these SNPs turn out to be dingo specific when tested on reference samples, then custom SNP genotyping would allow analysis for dingo or domestic dog origin. The average spacing between highly informative markers would be about 3 Mb.

The combination of next generation sequencing and SNP array genotyping technologies will allow unprecedented accuracy of amount and distribution of introgression of an invasive species. Not only will it ensure accurate testing, but it will also allow studies of selection in hybrids for retention of dingo or domestic dog loci.

SNP genotype testing is expected to enable quantification of genetic variation in dingoes, but care needs to be taken in selecting appropriate domestic dog controls. First,

27 the limited reference material currently used to define dingo alleles has come from captive populations, which may not be representative of all of the genetic variation across Australia. There is also a small possibility that some reference dingoes do have some undetected low levels of dog ancestry, caused by historical hybridisation events.

Second, the domestic dogs used as reference samples may not be representative of the breeds hybridising with dingoes in the wild. Some Australian dog breeds, notably

Australian Cattle Dog and Kelpies, are thought to have the dingo as one of their founding breeds. These breeds should be preferentially included in future studies. Third, some opponents to the use of genetic testing methods in the study of dingo-domestic dog introgression have suggested a need to type pre-European samples of dingoes as a non-hybrid reference population. Sequencing of mummified specimens may be possible in the future using ancient DNA, as next generation sequencing has already been applied to the mammoth (Miller et al. 2008). Alternatively, samples from remote isolated island populations may be effective reference alternatives to time-consuming and expensive ancient DNA techniques.

28 Table 1-2 Differences between 454 dingo sequence and Boxer reference sequence on chromosome 1 (5% of the dog genome) Differences of dingo from the dog Number of differences reference on Chromosome 1 Total* 4037 SNPs 1222 1 bp indels 2915 Indels > 1bp 150 *SNPs with reading of N (A, G, C or T) removed

Figure 1-4 SNP Distribution. Distribution of 328 SNPs that differ between the dog genome reference (v2.1) and 454 dingo sequence reads.

All 328 sites have high quality scores (> 20) in 454 sequence reads. Chromosome 1 (CFA1) accounts for 5% of the canine genome. SNPs

were identified by comparing dingo 454 sequence reads to the dog reference genome (v2.1).

1.9 Conclusion

The dingo is an ancient native dog that has been present in Australia for approximately

5,000 years. Dingoes are Australia’s top terrestrial predator and they play an important role in indigenous culture and spiritualism. Since Europeans colonised Australia in 1788 the genetic identity of the dingo has been threatened by hybridisation with domestic dogs. Hybrids are now widespread throughout Australia, but particularly in southeastern

Australia. Physical techniques including skull morphometrics are unable to distinguish between living dingoes and hybrids, and genetic methods must be utilised.

Microsatellites have been used since the 1990s to identify dingoes; however, this method has accuracy issues if the hybridisation event occurred more than 4 generations in the past. Investigation of coat colour and texture genetics will inform human perceptions of the dingo. When these studies are linked with SNP genotyping the genetic variation within dingoes can be more fully quantified. As personal genomics becomes available, SNP genotyping equipment may be developed that will allow genotyping in the field (Cyranoski 2005). This would be one of the most useful outcomes for conservation agencies, because it would allow quick assessment of the extent of hybridisation of an animal in the field.

The future of genetic testing is promising, but it is not enough to save the genetic identity of the dingo. Significant changes to legislation and control practices are required to protect dingoes from extinction in the wild. Unless major changes in policy are made, the only hope for the survival of the dingo is the establishment of captive

31 colonies for conservation, such as by the Native Dog Conservation Society at Bargo

NSW, zoos and wildlife parks. Perhaps, isolation of non-hybrid populations by fencing or natural isolation, such as the case for Fraser Island, is required to maintain the genetic identity of the dingo in the wild. Control of domestic dogs in sensitive dingo areas may also help. Unfortunately, government agencies are slow to make policy changes based on genetic evidence (Butchart et al. 2010).

32 Chapter 2

New insights on the history of canids in Oceania based on mitochondrial and

nuclear data

33 Abstract

How and when dingoes arrived in Oceania poses a fascinating question for scientists with interest in the historical movements of humans and dogs. The dingo holds a unique position as top terrestrial predator of Australia and exists in a wild state. In the first geographical survey of genetic diversity in the dingo using whole mitochondrial genomes, we analysed 16,428 bp in 25 individuals from five separate populations. We also investigated thirteen nuclear loci to compare with the mitochondrial population history patterns. Phylogenetic analyses based upon mitochondrial DNA and nuclear

DNA support the hypothesis that there are at least two distinct populations of dingo, one of which occurs in the northwest and the other in the southeast of the continent. The non-random geographical distribution of these divergent lineages suggests that plausibly there were two independent dingo introductions. Additionally the close relationship of the lineage found in southeastern Australia with the New Guinea Singing

Dog, supports the hypothesis that the split between the two dingo lineages occurred outside Australia. Molecular dating based upon the mitochondrial DNA suggests that the two independent mitochondrial lineages of dingoes diverged approximately 25,473 years before present. The observation of multiple dingo populations suggests the need for revision of existing conservation and management programs that treat dingoes as a single homogeneous population.

34 2.1 Introduction

Phylogeographic studies aim to relate patterns of genetic differentiation within and between populations to their geographic distribution and historical movements

(Hickerson et al. 2010; Wiens & Donoghue 2004). Such research can elucidate contemporary and historical processes, including geological landmass movements, climate change, sea level changes and commensal human movements, which drive differentiation and ultimately speciation (Hewitt 2000; Hickerson et al. 2010; Kokko &

López-Sepulcre 2006; Wiens 2004; Wiens & Donoghue 2004). A goal of this study is to investigate hypotheses concerning the origin, history and population structure of the

Australian native dog, the dingo.

Large carnivores, such as the dingo, are fundamental to the resilience of biodiversity and ecosystem functioning, yet face extinction due to conflict with humans over domesticated livestock (Ripple et al. 2014). The dingo is listed as vulnerable

(decreasing) on the IUCN red list, only extending across 84% of its historical range and threatened by hybridisation with European domestic dogs (Ripple et al. 2014). In other large carnivores, genetic subdivision has been linked to differences in habitat dispersal preference (Sacks et al. 2004, 2008), prey specialisation (Carmichael et al. 2001) and environmental conditions (Rueness et al. 2003a, 2003b; Stenseth et al. 2004). Despite this, there is a lack of scientific research concerning population size, genetic integrity and biogeography of dingoes. Subsequently ongoing management and control programs are not based on adequate and rigorous scientific data. Understanding the origin, history

35 and modern population structure of large carnivores is important for establishing adequate conservation strategies.

In Australia there is extensive historical and current conflict between livestock holders and dingoes. Dingoes are extensively managed and culled throughout Australia, but specifically in “prime” sheep country such as rural Queensland, New South Wales and

Victoria (Fleming et al. 2001). Legislatively, the dingo is considered to be both a native animal requiring protection and a controlled agricultural pest (Davis 2001; Downward

& Bromell 1990; Fleming et al. 2001, 2006). This dichotomy is largely due to lack of recognition of the dingoes’ ecological importance both as an apex predator and in controlling mesopredators such as foxes and cats (Brook et al. 2012; Carthey & Banks

2012; Fillios et al. 2010; Johnson et al. 2007; Johnson & VanDerWal 2009; Letnic et al.

2009a, 2009b, 2010, 2012b; Letnic & Dworjanyn 2011; May & Norton 1996; Moseby et al. 2012; Pople et al. 2000; Wallach et al. 2009a, 2009b, 2010). The dingoes’

‘modern’ arrival with humans (Savolainen et al. 2004) and a belief that they were responsible for the extinction of the Thylacine and Tasmanian Devil on mainland

Australia (Fillios et al. 2012; Letnic et al. 2012a) also impacts the public’s perception.

Interestingly, recent modelling suggests human population growth and climate change is the best explanation for the extinction of the Thylacine and Tasmanian Devil on mainland Australia rather than the often-blamed dingo (Roberts 2014).

Another important aspect of the dingo is its association with indigenous Australians and history of long separation from the global dog population until European colonisation in

1788 (Ardalan et al. 2012; Oskarsson et al. 2011; Savolainen et al. 2004; Smith &

36 Litchfield 2009). It has long been thought that dingoes (specifically their arrival time and route) may hold the key to early human interactions in and around Oceania

(Ardalan et al. 2012; Oskarsson et al. 2011; Sacks et al. 2013).

Ongoing controversy surrounds the origin of dogs (Pang et al. 2009; vonHoldt et al.

2010). Most hypotheses suggest that dogs in South East Asia and Polynesia likely radiated with humans but the timing of this event is debated (Pang et al. 2009; Sacks et al. 2013; vonHoldt et al. 2010). One hypothesis based on single nucleotide polymorphism (SNP) data suggests that the primary site of canine domestication was the Middle East, with only a few ancient dogs, such as the dingo, sharing affinity with

Asian wolves (vonHoldt et al. 2010). A Y chromosome based hypothesis suggests that a

Neolithic expansion of dogs from Southeast Asia partially replaced older dogs in the west and north (Sacks et al. 2013). An alternate hypothesis based on mitochondrial

DNA (mtDNA) data proposes the primary dog domestication site as Southern China about 16,000 years before present (BP) (Pang et al. 2009). However, complete mtDNA genomes of ancient canids from Eurasia suggest that all modern dogs are phylogenetically related to ancient or modern canids from Europe and molecular dating suggests domestication of dogs began 18,800-32,100 years BP (Thalmann et al. 2013).

Data from a 33,000 year old “Altai dog” skull found in southern Siberia revealed a mtDNA haplotype more closely related to modern dogs and prehistoric new world canids than to contemporary wolves; this study suggests dogs may have diverged from wolves more than 33,000 years ago (Druzhkova et al. 2013).

37 The species classification of dingoes is extremely controversial. Most studies suggest that dingoes fit within the dog phylogeny, and indicate a close relationship between dogs and wolves, i.e. Canis lupus familiaris or Canis lupus dingo (Ardalan et al. 2012;

Brown et al. 2011; Oskarsson et al. 2011; Pang et al. 2009; Sacks et al. 2013;

Savolainen et al. 2004; vonHoldt et al. 2010). Recently, a push has been made to classify the dingo as a separate species, Canis dingo, given its divergent morphology and long-term reproductive isolation (Crowther et al. 2014). This is logically consistent with the taxonomic naming of the New Guinea Singing Dog (NGSD) as Canis hallistromi (Koler-Matznick et al. 2004). Regardless, the dingo represents an important population of canids whose evolutionary history and biogeography may shed light on the evolution of domestic dogs and the relationship between humans and dogs in

Oceania.

It is often argued that dingoes were brought to Australia as a result of human movements (Ardalan et al. 2011; Corbett 1995; Macintosh 1964, 1975; Savolainen et al. 2004). There are two major pre-European human movements that occurred in the

Australasian region: (1) Indigenous Australians settled Australia in the Pleistocene, between 46,000-60,000 years BP (Gibbons 2001; McEvoy et al. 2010; Milham &

Thompson 1976; van Holst Pellekaan 2001, 2011), and (2) Neolithic expansion occurred in the Holocene, approximately 5,500 years BP (Bocquet-Appel 2011;

Diamond 2002; Kayser et al. 2001; Zheng et al. 2011). In the Pleistocene, Australia and

Papua New Guinea were joined by land bridges forming the continent of Sahul.

Approximately 6,500-8,000 years BP the land bridge between Papua New Guinea and

Australia opened, forming two geographically distinct islands. The Neolithic expansion

38 was responsible for replacing many hunter-gatherer cultures and introducing agriculture and domesticated animals, including pigs and chickens, to South East Asia and Oceania but not Australia (Bocquet-Appel 2011; Diamond 2002; Gibbons 2001; Haak et al.

2010; Larson et al. 2007, 2010).

The oldest fossilised dingo remains, excavated in southern Australia, have been dated to

3,500 years BP, suggesting an arrival date for dingoes of 5,000 years BP during the

Neolithic expansion (Gollan 1984; Gould 1863; Macintosh 1975; Mulvaney 1975).

Fossil evidence provides a solid minimum divergence estimate, if the fossil can be accurately identified. However, the fossil record is patchy and so divergence estimates may be underestimates. This is important in Island South East Asia and northern

Australia as acidic soil and tropical conditions have likely inhibited fossil recruitment and persistence (Behrensmeyer 1978; Kidwell & Flessa 1996; Tappen 1994). The possibility that dingoes have been in Australia longer than previously thought is supported by a recently discovered rock art painting identified as a dingo by elders of the region (David et al. 2013). The painting is located at Nawarla Gabarnmang

(Northern Territory, Australia), where an excavated painted rock from the remote site was recently radiocarbon dated to 28,000 years BP (David et al. 2013). At this time it is not clear why the estimates from the fossil record and rock art painting differ. One possibility is that hitherto undiscovered fossils may exist in northern Australia. An alternate possibility is that the rock art painting is not a dingo but perhaps a Thylacine or other marsupial. In this study we employ molecular data to study population genetic subdivision in the Australian dingo and estimate their time of divergence.

39 Molecular data has previously been used to unravel geographic subdivision and estimate the time to the most recent common ancestor (MRCA) in a variety of animals including dingoes. Dating divergence times typically rely upon estimating a substitution rate using sequence data along with biogeographic, pedigree, geological or archaeological evidence. The initial genetic study on dingoes using the mtDNA control region estimated the time to the MRCA at approximately 5,000 years BP (Savolainen et al.

2004). A more extensive study of dingoes and South East Asian island dogs re- estimated the divergence of the mtDNA to 4,600-18,300 years BP using the same control region (Oskarsson et al. 2011). A Y chromosome study on 47 wild and captive dingoes largely from eastern Australia identified a pattern of West versus East biased allele distribution and suggests that dingoes may have immigrated twice (Ardalan et al.

2012). A close relationship to the NGSD was also inferred from Y chromosome data

(Ardalan et al. 2012; Sacks et al. 2013). There are several issues with these papers.

First, several of these studies sourced animals from captive breeding programs. In some circumstances this is unavoidable; however, it can make it difficult to interpret aspects of population subdivision if the geographic and genealogical history of the animals is not accurately documented.

A second issue is intraspecific rate heterogeneity. Intraspecific rate heterogeneity in specific genes is a significant issue with molecular dating where substitution rates are estimated using distant fossil or interspecific calibration points (Ho et al. 2011; Ho &

Larson 2006). Rate heterogeneity could lead to misleading divergence estimates, particularly if the timeline is less than 100,000 years (Ho et al. 2011; Ho & Larson

2006). Third, these studies do not test for selection and selection can bias estimates of

40 divergence time (Ballard & Whitlock 2004). A fourth more general issue is that the coalescent time of a particular gene may not reflect the divergence time of the taxa. This is particularly true with mtDNA data, though it is also a concern for Y chromosome and autosomal data (Ballard & Whitlock 2004; Edwards & Beerli 2000; Humphries &

Winker 2011; Rubinoff & Holland 2005; Toews & Brelsford 2012).

The presence of genetic subdivision within the dingo has been postulated for decades but has yet to be empirically investigated (Corbett 1995, 2001; Savolainen et al. 2004).

We assess the phylogenetic relationships between five geographically distinct dingo populations, the NGSD and two closely related dogs from South East Asia. These five dingo populations enable us to investigate genetic variation spanning the Australian continent; previous studies have not sampled as broadly (spatially and ecologically).

Two dogs from Indonesia (Bali and Kalimantan) are included as plausible relatives of dingoes. All dingo samples were sourced from wild populations. We further investigate patterns of phylogeny and differentiation using whole mtDNA genome data and thirteen nuclear loci from three functional groups: the major histocompatibility complex (DLA) genes, olfactory receptor genes (cfOR) and coat colour genes (CC). We then cautiously employ the whole mtDNA genome data and the published canid mtDNA substitution rate to the estimate the divergence times of lineages. These estimates are older than previously reported. Based upon these data we hypothesise that there are at least two dingo populations and that plausibly dingoes immigrated into Australia more than once.

There is also evidence that dingoes were not brought to Australia as part of the

Neolithic expansion.

41 2.2 Materials and Methods

2.2.1 Canid sampling

2.2.1.1 Mitochondrial DNA

We sampled five dingoes from each of five geographically distinct regions: the

Kimberley in northwestern Australia, the Gibson and Simpson Deserts in western and central Australia respectively, Fraser Island in eastern Australia and the Australian

Alpine region in southeastern Australia (Appendix 2). Tissue and/or blood was collected and pre-screened for domestic dog introgression, using a microsatellite based test developed by ANW (Wilton 2001; Wilton et al. 1999). These microsatellites were selected to distinguish between domestic dogs and dingoes and are not informative of population subdivision in the dingo. Only wild dingoes, which were verified as pure by this method, were assigned to this project (Stephens 2011; Wilton 2001; Wilton et al.

1999).

Given the inferred close relationship of the dingo to the wild dog of Papua New Guinea we include a NGSD sample (Ardalan et al. 2012; Sacks et al. 2013). A single NGSD was used in whole mtDNA dataset because the captive NGSD in North America predominately derive from the same maternal lineage/founder (personal communication

J Koler-Matznick). The Papua New Guinea sample was collected from captive NGSD housed in the USA (Appendix 2). NGSD have not been observed in the wild since the

1950s. We also included dog samples from the islands of Bali and Kalimantan

(Indonesia) to investigate the relationship of dingoes to other dogs in the region

42 (Oskarsson et al. 2011; Savolainen et al. 2004). The Bali dog sample was collected from Bali, Indonesia and the Kalimantan dog sample was collected from Mallinau on the island of Kalimantan, Indonesia (Appendix 2).

2.2.1.2 Nuclear DNA

For nuclear DNA investigations we followed the same dingo sampling design as for mitochondrial investigations. However, due to difficulties amplifying nuclear markers the Fraser 1 and Simpson 4 individuals were replaced with samples from the same geographical region (Fraser 6 and Simpson 6). Further, we expected additional nuclear variation from the captive NGSD housed in the USA, so we increased the number of

NGSD samples to five. Finally, we had difficulty amplifying the nuclear markers from the Bali dog sample and it was subsequently excluded from all nuclear studies.

2.2.2 Genetic investigations

2.2.2.1 MtDNA genome analysis

To capture all the variation in the mtDNA we sequenced the complete genome. It has been observed that the mtDNA control region exhibits low variation in some canines and thus may underestimate divergence times (Pang et al. 2009; Webb & Allard 2009).

DNA was extracted using a Qiagen DNeasy kit (Qiagen Sciences, Germantown, USA) and the complete mitochondrial genome of 25 dingoes, 1 NGSD and 2 Indonesian canines were amplified in 33 overlapping PCR reactions (Appendix 3-1). PCR amplicons underwent ExoSAP-IT® (USB Amersham, Buckinghamshire, UK) purification prior to sequencing. Sanger sequencing was performed on the purified

43 templates using BigDye terminator v3.1 (Applied Biosystems Inc., Foster City, USA).

To test the potential for nuclear encoded mitochondrial fragments to influence our results we performed a rolling circle amplification on one sample (Wolff et al. 2012).

This protocol corroborated our results and suggested the data were not influenced by the amplification of nuclear encoded mitochondrial fragments.

Sequence chromatograms were visualised and manually edited to remove erroneous base calls, such as dye blobs, using Sequencher 5.1 (Gene Codes corp., Ann Arbor,

USA). Mutations observed in a single individual were corroborated by re-amplifying and sequencing the region of interest and comparing the independent chromatograms.

No instances of mismatch, damage or misincorporation were identified. MtDNA genome contigs were then constructed for each individual.

2.2.2.2 Nuclear gene analysis

Thirteen nuclear regions were identified as plausibly containing genetic variation in dingoes (Berryere et al. 2005; Cadieu et al. 2009; Candille et al. 2007; Dreger &

Schmutz 2010; Everts et al. 2000; Hong et al. 2009; Kennedy et al. 2007; Kerns et al.

2003, 2004, 2007; Newton et al. 2000; Robin et al. 2009; Runstadler et al. 2006;

Schmutz & Berryere 2007; Schmutz et al. 2003, 2007; Zhang et al. 2011). This included genes involved in immunity (DLA), olfactory reception (cfOR) and coat colour (CC). DNA was extracted using a Qiagen DNeasy kit (Qiagen Sciences,

Germantown, USA) and the thirteen nuclear loci were amplified using PCR in 25 dingoes, 5 NGSD and 1 Indonesian dog (Appendix 3-2). PCR amplicons underwent

ExoSAP-IT® (USB Amersham, Buckinghamshire, UK) purification prior to sequencing.

44 Sanger sequencing was performed on the purified templates using BigDye terminator v3.1 (Applied Biosystems Inc., Foster City, USA).

Sequence chromatograms were visualised and manually edited to remove erroneous base calls using Sequencher 5.1 (Gene Codes corp., Ann Arbor, USA). Heterozygous sequences were resolved in a two-step approach. First, we used the Phase v2.1 algorithm implemented in DnaSP (Librado & Rozas 2009) to identify individuals carrying unique alleles. Second each unique allele was cloned to fully resolve the sequence. Cloning was performed using a CloneJet kit according to the manufacturers instructions (Thermo Fisher Scientific Inc., Waltham, USA).

2.2.3 Phylogeny

2.2.3.1 Whole mtDNA phylogenetic analysis

According to the literature, the most suitable phylogenetically distinct outgroup taxa for a critical examination of dingoes are South East Asian dogs, most specifically those carrying the A29 control region haplotype (Oskarsson et al. 2011; Savolainen et al.

2004). Therefore we chose A29_10100 (GenBank # EU789776), from Guizhou in southern China and two Indonesian dogs from the islands of the Bali and Kalimantan as outgroups. These dogs are phylogenetically distinct from dingoes and NGSD.

The algorithms employed in JModeltest 2.3.1 (Darriba et al. 2012) allow scientists to select the most suitable phylogenetic substitution model for a given DNA sequence alignment. JModeltest 2.3.1 (Darriba et al. 2012) was run on the dataset of 25 dingoes,

45 NGSD, Bali, Kalimantan and A29_10100 dogs to evaluate the most suitable substitution model for the dataset included in this study. A GTR + G + I substitution model was chosen using the Akaike selection criteria, and used in all subsequent mtDNA phylogenetic analysis. Constant population size coalescent models are generally most suitable for intraspecific datasets (Kingman 1982). Thus Bayesian phylogenetic analyses were performed in BEAST v1.7.4 (Drummond et al. 2012) with a constant population size coalescent model.

A strict clock is suitable for analyses that incorporate intraspecific data (Brown & Yang

2011); therefore mtDNA phylogenetic analyses were run with a strict molecular clock

(normally distributed with mean = 1.28 x 10-8 mutations-1 site-1 year-1 and stdev = 3.27 x

10-9) (Pang et al. 2009). All runs were optimised for MCMC chain steps to ensure the estimated sampling size was above 200 in Tracer 1.5 (Rambaut & Drummond 2007).

Sampling occurred every 5,000 steps with a 10% burn-in. A simple chi-squared contingency table analysis was performed to investigate whether the observed geographic subdivision departs from a random pattern.

2.2.3.2 Nuclear phylogenetic analysis

We chose a “total evidence” approach to the analysis of our nuclear gene dataset. Hence instead of constructing single gene trees we concatenated the autosomal gene fragments for phylogenetic analysis. Concatenation is particularly useful in situations where it is thought that divergences are recent, gene sequences are short or gene regions are highly conserved and carry few substitutions (Gadagkar 2005). However, concatenation may introduce bias through genetic signal swamping or rate heterogeneity differences

46 between genetic loci (Edwards 2009). Therefore analyses based upon concatenated gene sequences need to be treated with a degree of caution, as they may not depict the only plausible phylogeny.

Jmodeltest 2.3.1 (Darriba et al. 2012) was run on the nuclear dataset to determine the most suitable substitution model. Substitution models were chosen with the Akaike selection criteria. For the concatenated analyses of 13 nuclear loci heterozygote sites were assigned using the IUPAC code. The concatenated analyses were run with a GTR

+ G + I substitution model, as chosen by JModelTest 2.3.1 (Darriba et al. 2012). All

Bayesian phylogenetic analyses were performed using BEAST v1.7.4 (Drummond et al.

2012) with a constant population size coalescent model. All runs were optimised for

MCMC chain steps to ensure the estimated sampling size was above 200 in Tracer 1.5

(Rambaut & Drummond 2007). Sampling occurred every 5,000 steps with a 10% burn- in. No clock was enforced for nuclear analyses. Simple chi-squared contingency table analyses were performed to investigate whether the observed allele frequency pattern at the cfOR0007 and cfOR0034 cfOR gene regions departed from a random pattern indicating geographic subdivision.

2.2.3.3 Topology testing

To investigate whether dingoes form a monophyletic lineage the Bayesian analyses for both the mitochondrial and nuclear datasets were run a second time with a constrained topology (i.e. dingoes forced to be monophyletic). The marginal likelihood of the unconstrained and constrained phylogenies was calculated in Tracer v1.5 (Rambaut &

Drummond 2007) using 2,500 bootstrap replicates. Bayes factors were calculated in

47 Tracer v1.5 (Rambaut & Drummond 2007) using the following formula:

2log10(BF)=2log10(L(H0)) – 2log10(L(HA)) where H0 is the optimal unconstrained topology and HA is the constrained tree topology. Bayes factors were interpreted using the guidelines described by Kass & Raferty (1995). Briefly, values between 2 and 6 indicate positive evidence against HA, 6 to 10 strong evidence against HA and higher than 10 very strong evidence against HA (Kass & Raferty 1995).

2.2.4 Statistical analyses

2.2.4.1 MtDNA

Indices of nuclear diversity and variability, π and θ, were calculated for the complete mtDNA, coding and RNA mtDNA regions and the 582 bp mitochondrial control region in DnaSP v5.10.1 (Librado & Rozas 2009). Sliding window analysis was carried out in

DnaSP v5.10.1 (Librado & Rozas 2009) using a window of 200 bp and steps of 50 bp.

To investigate if the sampled mtDNA sequences are evolving in a manner consistent with a strictly neutral equilibrium pattern of evolution Tajima’s D (Tajima 1989), Fu and Li’s F* and Fu and Li’s D* tests were performed using DnaSP v5.10.1 (Librado &

Rozas 2009). Neutrality tests such as Tajima’s D, Fu and Li’s F* and Fu and Li’s D* compound the effects of demography and selection, therefore selection can really only be reliably identified when the population is at equilibrium (Fu 1997; Fu & Li 1993;

Tajima 1989). If a departure from demographic equilibrium is observed then significant values should be evaluated with caution (Fu 1997; Fu & Li 1993; Tajima 1989). To avoid violating the assumptions of the Neutrality tests, such as the absence of population subdivision, statistics were calculated on phylogenetic lineages separately.

48 Measures of genetic differentiation (FST) between geographical populations (Alpine,

Fraser, Simpson, Gibson, Kimberley, Papua New Guinea and Indonesia) were calculated in Arlequin 3.5 (Excoffier & Lischer 2010). FST calculations are frequently used in population genetics to investigate whether populations of the same species are subdivided or still experiencing high levels of gene flow. Wright (1978) deﬁned

‘‘moderate differentiation’’ as being FST values between 0.05 and 0.15; values above

0.25 would be considered ‘‘very great’’ genetic differentiation and values less than 0.05 were deﬁned as evidencing low levels of genetic differentiation and ‘‘higher’’ levels of gene ﬂow.

An Analysis of Molecular Variance (AMOVA) to test for the genetic structure was carried out in Arlequin 3.5 (Excoffier & Lischer 2010). Groups were defined as identified in the mtDNA phylogenetic analyses. An AMOVA is used to statistically evaluate the amount of population genetic structure observed at a genetic locus

(Excoffier et al. 1992).

2.2.4.2 Nuclear

Indices of nuclear diversity and variability, π and θ, were calculated for each autosomal locus using the fully resolved datasets in DnaSP v5.10.1 (Librado & Rozas 2009).

Tajima’s D (Tajima 1989), Fu and Li’s F* and Fu and Li’s D* (Fu 1997; Fu & Li 1993) tests were performed using DnaSP v5.10.1 (Librado & Rozas 2009). Neutrality statistics were calculated for genetic lineages, as defined by the mitochondrial phylogeny, separately. FST and AMOVA analyses were implemented in Arlequin 3.5 (Excoffier &

Lischer 2010).

49 2.2.5 Estimating divergence time and substitution rate of mtDNA

We estimated divergence times using 28 whole mtDNA genomes and the published estimate of the whole mtDNA substitution rate of canids (Pang et al. 2009), which was calibrated using the coyote-wolf fossil divergence time of 1.5 - 4.5 million years BP

(Nowak 2003). This is 1.28 x 10-8 mutations-1 site-1 year-1 with a standard deviation of

3.27 x 10-9 (Pang et al. 2009). We calculated molecular divergence by implementing a tMRCA statistic for each divergence event observed, running the analysis in Beast v1.7.4 (Drummond et al. 2012) and visualising the results in Tracer 1.5 (Rambaut &

Drummond 2007).

In parallel, we followed Oskarrson et al. (2011) and constructed maximum parsimony networks using TCS v1.21 (Clement et al. 2000). We then calculated the average number of mutational steps, ρ, from any given individual of a lineage to the hypothesised MRCA. Analysis was run treating gaps as missing data and assuming a

95% connection limit. To estimate the divergence time range from the hypothesised

MRCA, ρ was multiplied by the mtDNA substitution rate. A potential problem with these approaches is that they may overestimate the intraspecific divergence time because slightly deleterious mutations can accumulate within species or populations but not go to fixation between them (Ho et al. 2011; Ho & Larson 2006).

50 2.3 Results

2.3.1 Genetic analyses

2.3.1.1 MtDNA genome analysis

The dingo, NGSD, Bali and Kalimantan dog mitochondrial genome architecture is identical to that of other dogs. Excluding the repeat structure located within the control region that is difficult to unambiguously align, the mtDNA genome analysis included

16,428 bp.

Among the 25 dingo mtDNA genomes sequenced, 20 haplotypes were detected. The

NGSD, Bali and Kalimantan dogs each carried distinct mitochondrial haplotypes. In the

20 dingo and 1 NGSD haplotypes observed there were 78 segregating sites (excluding indels) and 17 non-synonymous polymorphisms (Table 2-1). A single mitochondrial haplotype was detected in all of the dingoes from Fraser Island. This is consistent with a small maternal foundation population.

2.3.1.2 Nuclear gene analysis

Coding regions of 13 nuclear loci were sequenced. There were a total of 87 segregating sites observed at the three DLA genes, 127 at the seven cfOR genes and 14 from the three CC genes (Table 2-1).

51 2.3.2 Phylogeny

2.3.2.1 Whole mtDNA phylogenetic analysis

Bayesian analyses run on the 28 genomes sequenced and the Chinese A29_10100 dog corroborate preliminary analyses suggesting that the Bali, Kalimantan and A29_10100 dogs sit outside the monophyletic NGSD and dingo lineage (Figure 2-1).

A notable result of this phylogenetic analysis is the observation of strong biogeographic clustering of two lineages, with a southeastern (SE) population and predominantly northwestern (NW) population. Due to the low posterior probability support of 0.4 it is not clear, however, whether the NGSD is distinct from the SE lineage of dingoes. The single exception to the biogeographic clustering is that one dingo (Alpine 5) collected from the Australian Alpine region in the southeast of Australia was found to carry a NW mtDNA lineage haplotype (Figure 2-1). A simple contingency table analysis comparing the distribution of the two lineages within the five geographic regions confirms that the

2 geographic clustering is non-random (χ 4= 21.4; p=0.000).

Table 2-1 Descriptive measures of Nucleotide diversity for mtDNA and nuclear gene regions Function Gene N* bp S h Hd All 26 16428 78 21 0.97 8.40 x 10-04 1.24 x 10-03 MtDNA Coding and RNA's 26 15819 72 21 0.97 8.20 x 10-04 1.19 x 10-03 Control region 26 582 6 7 0.67 1.42 x 10-03 2.70 x 10-03 CBD103 60 344 2 3 0.13 3.80 x 10-04 1.25 x 10-03 Coat Colour ASIP 60 487 3 3 0.26 1.03 x 10-03 1.32 x 10-03 MC1R 60 1102 7 8 0.71 1.10 x 10-03 1.61 x 10-03 cfOR0007 60 785 16 2 0.35 7.04 x 10-03 4.37 x 10-03 cfOR0011 60 837 9 7 0.43 2.04 x 10-03 2.31 x 10-03 cfOR0034 60 866 15 2 0.49 8.56 x 10-03 3.71 x 10-03 Olfactory -02 -02 Receptor cfOR0123 60 784 47 27 0.87 1.27 x 10 1.34 x 10 cfOR0184 60 866 3 4 0.38 4.80 x 10-04 7.40 x 10-04 cfOR14A11 60 781 18 9 0.33 2.55 x 10-03 4.94 x 10-03 DOPRH07 60 857 19 4 0.13 2.59 x 10-03 4.75 x 10-03 DQA1 60 292 9 5 0.64 9.23 x 10-03 6.61 x 10-03 DLA DQB1 60 264 37 10 0.72 4.69 x 10-02 3.41 x 10-02 DRB1 60 268 41 14 0.81 6.81 x 10-02 4.00 x 10-02 * Dingo and NGSD samples pooled for calculation of descriptive nucleotide statistics given close phylogenetic relationship. Column headings are defined as follows bp: base pairs; S: segregating sites; H: number of haplotypes; Hd: haplotype diversity; : nucleotide diversity and : theta per site.

Alpine 1 Alpine 3 1 Alpine 2 1 Alpine 4 Fraser 2 SE 1 Fraser 5 0.4 Fraser 1 Fraser 4 Fraser 3 NGSD 1 Gibson 2 0.7 0.99 Gibson 4 Alpine 5 Simpson 1 Kimberley 1 1 1 Kimberley 3 Kimberley 2 Gibson 1 1 Kimberley 5 Kimberley 4 NW 0.89 Simpson 3 Simpson 4 Simpson 2 1 Simpson 5 1 Gibson 5 Gibson 3 1 Kalimantan Bali A29_10100 20000.0

Figure 2-1 Bayesian analysis of 29 canine mtDNA genomes. Phylogram constructed using a GTR+G+I model in BEAST v1.7.4 (Drummond et al. 2012), assuming a strict clock and a coalescent model with constant population size. Posterior probability values are reported above nodes and values less than 0.6 are not shown, except for the

NGSD/SE lineage node that is of particular interest. Italics indicate samples with discordant mtDNA and nuclear DNA phylogenies. The scale bar indicates units of time estimated using a substitution rate of 1.28 x 10-8 mutations-1 site-1 year-1 with a standard deviation of 3.27 x 10-9 (Pang et al. 2009).

54 2.3.2.2 Nuclear phylogenetic analysis

We performed phylogenetic analyses by concatenating all the nuclear genes. The concatenated analysis observed biogeographic clustering of the NW populations and

SE/NGSD populations with a posterior probability support of 0.99 (Figure 2-2). This phylogeny also corroborates the close relationship between the NGSD and SE lineage with posterior probability support of 0.99 (Figure 2-2). There are four exceptions to the strict biogeographic clustering of the NW and SE lineages. First, like the mtDNA results the Alpine 5 individual continues to cluster with the NW lineage. Unlike the mtDNA results; the Fraser 3 and Fraser 4 dingoes group with the NW lineage, Simpson

5 now clusters outside the SE/NGSD group, while Alpine 3 and Alpine 4 cluster closer to the Indonesian dog (albeit with a low level of posterior probability, Figure 2-2).

A contingency table analysis, comparing the distribution of the two divergent nuclear

2 alleles between the two mtDNA lineages, at cfOR0007 (χ 2= 11.7; p=0.003) and

2 cfOR0034 (χ 2= 6.19; p=0.045) indicates that there is a non-random distribution of these olfactory nuclear haplotypes between the two dingo mtDNA lineages. This suggests that the observed nuclear subdivision is highly correlated with that detected in the mtDNA.

55 Alpine 4 Alpine 3 0.61 Alpine 2 0.72 Alpine 1 0.99 Fraser 5 SE Fraser 2 0.99 0.95 Fraser 6 NGSD 3 1 NGSD 5 0.99 NGSD 1 NGSD NGSD 4 NGSD 2 Simpson 5 Fraser 3 Fraser 4 Simpson 6 Gibson 3 0.99 Gibson 1 Gibson 2 Simpson 3 0.99 Simpson 1 Kimberley 2 NW Gibson 5 Simpson 2 0.99 Gibson 4 Kimberley 3 Kimberley 4 0.95 Kimberley 5 Kimberley 1 Alpine 5 Kalimantan

0.002

Figure 2-2 Bayesian analysis of 31 canine concatenated nuclear haplotypes. Phylogram constructed using a GTR+G+I model and a coalescent model with constant population size in BEAST v1.7.4 (Drummond et al. 2012). Posterior probability values are reported above nodes and values less than 0.6 are not shown. Italics indicate samples with discordant mtDNA and nuclear DNA phylogenies. The scale bar indicates an estimate of the average number of substitutions per site between two nodes.

56 2.3.2.3 Topology testing

Topology testing was carried out to investigate whether dingoes were non-monophyletic as suggested by unconstrained mitochondrial and nuclear phylogenies (Figure 2-1 and

Figure 2-2). In the topology analyses the null hypothesis or H0 was that dingoes are non-monophyletic, as depicted in the unconstrained phylogenies (Figure 2-1 and Figure

2-2). The alternative hypothesis or HA was that dingoes are monophyletic, as enforced in the constrained phylogenies (not shown). Log10 Bayes factor results indicated that there was very strong evidence against monophyly of dingoes (i.e. HA was rejected), for both the mitochondrial and nuclear datasets (Table 2-2).

2.3.3 Statistical analyses

2.3.3.1 MtDNA

For the whole mtDNA data π was 8.4 x 10-4 and θ was 1.24 x 10-3. Values for the coding, tRNA and rRNA mtDNA regions were similar (Table 2-1). As recently reported for dogs (Webb & Allard 2009), a sliding window analysis of the genomes shows genetic variability in the mitochondrial control region was low (Figure 2-3). Indeed, π was 1.42 x 10-3 and θ was 2.7 x 10-3 (Table 2-1).

57 Table 2-2 Topology testing to investigate the monophyly of dingoes using log10 Bayes factors Unconstrained topology Constrained* (HA) Bayes (H0) marginal likelihood marginal likelihood factor Mitochondrial dataset -23193.80 -23222.25 12.53(*) Nuclear dataset -14066.99 -14613.59 237.39(***) *Topology constrained to dingoes being monophyletic - Bayes factors interpreted using Kass and Raferty (1995) with positive (*), strong (**) and very strong (***) evidence against HA.

Figure 2-3 Sliding Window Analysis of Dingo mitochondrial genomes. Sliding window of 200 nucleotides with steps of 50 nucleotides. Sites with gaps were excluded. The grey bar indicates the mitochondrial control region.

59 We calculated Tajima’s D for all dingoes and for each mtDNA lineage. Negative neutrality test values suggest the presence of ongoing demographic processes such as range expansion and/or population expansion or purifying selection (Fu & Li 1993;

Tajima 1989). Positive neutrality tests values, on the other hand, suggest demographic processes such as population decline or range constriction or balancing selection (Fu &

Li 1993; Tajima 1989). When measures of neutrality were calculated for each lineage, the SE lineage had a positive but non-significant Tajima’s D value whilst the NW lineage had a negative but non-significant Tajima’s D value (Table 2-3). Fu and Li’s F* and D* results were not significant but did corroborate the Tajima’s D findings (Table

2-3). The differing test values between the mtDNA lineages provide initial evidence that the lineages may have been subject to different demographic processes.

FST values corroborate the phylogenetic pattern of differentiation observed, with a SE lineage and NW lineage; however, values are high between the Fraser Island and Alpine populations (Table 2-4). This is expected because Fraser Island is geographically isolated. FST values tended to be slightly lower between phylogenetically related populations (Table 2-4).

AMOVA results further corroborate the presence of population subdivision within dingoes, both at the lineage level and population level (Table 2-5). Indeed, AMOVA results were significant among groups and among groups within populations.

60 Table 2-3 Neutrality test results for mitochondrial and nuclear gene regions NW Lineage NGSD SE Lineage Tajima's Fu&Li's Fu & Li's Tajima's Fu&Li's Fu & Li's Tajima's Fu&Li's Fu & Li's Gene D D* F* D D* F* D D* F* MtDNA -1.68 -1.99 -2.21 - - - 1.16 0.67 0.88 Average (Nuclear DNA) -0.32 -0.37 -0.35 -0.9 -1.07 -1.03 0.29 0.68 0.65 DQA1 -0.02 0.03 0.02 - - - -0.36 1.38 (*) 1.03 DLA DQB1 -1.08 -1.18 -1.36 - - - 0.2 1.67 (**) 1.44 (#) DRB1 0.55 1.18 1.14 - - - 0.91 1.68 (**) 1.69 (*) 3.18 cfOR0007 1.57 (**) 2.45 (**) -2.01 (**) -2.36 (**) -2.56 (**) - - - (***) cfOR0011 -1.72 (#) -0.61 -1.11 -1.74 (*) -2.01 (#) -2.18 (#) 1.3 1.35 (#) 1.54 cfOR0034 2.61(***) 1.55 (**) 2.22 (**) - - - 1.75 (#) 1.51 (**) 1.83 (**) Olfactory cfOR0123 -0.59 0.5 0.17 -0.64 -0.00039 -0.18 -0.71 0.54 0.2 Receptor cfOR0184 -0.63 -1.51 -1.45 - - - 0.49 0.67 0.71 -2.34 cfOR14A11 -4.08 (**) -4.14 (**) -1.56 (#) -1.78 -1.93 -0.46 -0.97 -0.95 (***) DOPRH07 -0.54 1.29 0.83 ------CBD103 -1.05 -0.76 -0.1 ------Coat ASIP -0.72 -0.28 -0.47 - - - 0.63 0.88 0.94 Colour MC1R -0.51 -0.85 -0.87 1.46 0.8 1.69 -1.74 (#) -1.92 -2.15 - Indicates neutrality statistics could not be calculated as no polymorphisms were observed within the population/lineage at this gene region. (*) indicates p<0.05 (**) indicates p<0.02 (***) indicates p<0.001 (#) indicates 0.10

Table 2-4 FST values for MtDNA (bottom) and nuclear gene regions (averaged) (top) Alpine Fraser Gibson Kimberley Simpson NGSD Indonesia Alpine - 0.14 0.23 0.26 0.19 0.21 0.24 Fraser 0.57 - 0.20 0.31 0.13 0.17 0.44 Gibson 0.43 0.75 - 0.06 0.01 0.32 0.42 Kimberley 0.49 0.82 0.17 - 0.12 0.39 0.47 Simpson 0.56 0.88 0.11 0.30 - 0.29 0.36 NGSD 0.46 1.00 0.54 0.67 0.77 - 0.46 Indonesia 0.51 0.86 0.57 0.65 0.72 0.42 -

Table 2-5 AMOVA# results for mtDNA and nuclear loci Markers Among Groups Among Populations Within Groups Among Individuals within Populations Within Individuals mtDNA 49.01 (***) 16.69 (***) 34.30 (***) - DQA1 40.2 (*) 6.7 14.7 38.4 (***) DLA DQB1 47.3 (**) 6.1 15.0 (***) 31.6 (***) DRB1 46.6 (***) 4.3 11.2 37.9 (***) cfOR0007 32.7 (*) 3.5 -4.8 68.5 cfOR0011 22.2 -0.1 33.2 (***) 44.6 (***) cfOR0034 24.1 27.07 (***) -2.1 50.9 (*) Olfactory cfOR0123 18.8 (*) -2.7 -3.5 87.4 Receptor cfOR0184 0.8 15.6 -1.3 85 cfOR14A11 20.2 -4 32.7 51.1 (**) DPORHO7 6.5 34.2 17.4 41.9 (***) CBD103 -4.2 -1.8 1.7 104.3 Coat ASIP -13.4 33.0 (*) -5.4 85.8 Colour MC1R 21.8 8.4 -0.6 70.4 (*) #AMOVA testing the following genetic structure: Group 1 (Alpine & Fraser), Group 2 (Simpson, Kimberley and Gibson), Group 3 (NGSD) and Group 4 (Indonesia) (*) indicates p <0.05 (**) indicates p <0.02 (***) indicates p <0.001

2.3.3.2 Nuclear

To investigate patterns of nuclear variation we examined nucleotide diversity and then conducted neutrality tests, FST calculations and AMOVA analyses. For the 13 nuclear loci, π ranged between 6.81 x 10-2 and 3.8 x 10-4, and θ ranged from 4.0 x 10-2 to 1.25 x

10-3 (Table 2-1). The gene regions with the highest π and θ values were DQB1 and

DRB1, and the lowest were cfOR0184 and CBD103 (Table 2-1).

We calculated Tajima’s D values based upon mitochondrial lineage (i.e. divided into SE versus NW lineages) and observed that there is a general trend for the SE lineage to have positive values, whilst the NW lineage has negative values (Table 2-2). Two particularly interesting genes that appear to be under selection in the NW lineage were cfOR0007 and cfOR0034. These olfactory genes showed significant positive Tajima’s

D, Fu and Li’s D* and Fu and Li’s F* values despite the general trend for the lineage to exhibit negative values (Table 2-3). All other results should be treated with caution, as the average values for the SE lineage are positive while those for the NW lineage are negative (Table 2-3).

FST values were calculated for each individual nuclear gene region and then averaged across all 13 regions to give a “multi locus” view (Table 2-4). These FST values corroborate the mtDNA, and nuclear gene phylogenies (Figure 2-1 and Figure 2-3). The multi locus FST values suggest a moderate level of differentiation between the Alpine and Fraser populations but great levels of differentiation between the southeastern populations and the Kimberley, Gibson and Simpson dingo populations. Similarly, the

Alpine and Fraser populations are slightly less divergent from the NGSD population

64 than the Kimberley, Gibson and Simpson populations according to the FST calculations.

All the dingo and NGSD populations are highly differentiated from the Indonesian dog based on FST values. The Simpson population appears to exhibit slightly lower FST values than the other NW lineage populations suggesting there is plausibly ongoing gene flow between the SE and NW populations in this region (Table 2-4).

AMOVA analyses were run to investigate patterns of geographic substructure (Table 2-

5). Significant among group variation was observed at the three DLA loci (DQA1,

DQB1 and DRB1) and at two OR loci (cfOR0007 and cfOR0123). Significant among population variation was observed at the OR cfOR0034 and the CC locus ASIP (Table

2-5).

2.3.4 Estimating divergence time & substitution rate of mtDNA

We estimated mtDNA divergence implementing a tMRCA statistic for each lineage using the substitution rate obtained from the dog whole mtDNA calibrated with the wolf-coyote fossil divergence by Pang et al. (2009) (Table 2-6). This substitution rate is likely an underestimate due to the ancient calibration point implemented by Pang et al.

(2009); as such divergence estimates should be regarded as maximum values. Under this scenario, conservative estimates of mtDNA divergence time between the SE lineage

(and NGSD) and the NW lineage of dingoes are at least 25,400 (25,473 – 93,806, 95%

HPD) years BP. A conservative divergence estimate of the NW dingo lineage is at least

12,300 (12,330 – 47,634, 95% HPD) years BP and the SE lineage is at least 9,700

(9,765 – 46,889, 95% HPD) years BP (Figure 2-1, Table 2-6).

In parallel, we calculated average pairwise differences within and between lineages following Oskarsson et al. (2011) (Table 2-6). Most generally, the estimated divergence time using this latter method is about 30% older than that estimated using Bayesian analysis. It is important to note that divergence estimates lower than 3,500 years conflict with the known minimum arrival time of the dingo set by the fossil record (Macintosh

1964).

Table 2-6 Estimates of divergence times of dingo lineages tMRCA tMRCA divergence Divergence Divergence divergence estimates (95% estimate estimates (95% estimate median3 HPD)3 1,2 median1,3 HPD)1,3 MRCA dingoes, NGSD & Indonesian dogs 68,790 30,380 107,200 11.5 89,231 27,885-150,577 MRCA dingoes & NGSD 59,640 25,473 93,806 10.16 81,280 25,400 - 137,160 MRCA NGSD and SE Lineage 55,522 22,723 88,321 7.80 62,400 19,500-105,300 MRCA NW Lineage 29,982 12,330 47,634 5.50 44,000 13,750 - 74,250 MRCA SE Lineage 28,327 9,765 46,889 4.78 38,222 11,944 - 64,500 1following Oskarsson et al. (2011). 2 is the average number of mutational steps from any given individual of a group / lineage to the hypothesised most recent common ancestor (MRCA). 3 years before present

2.4 Discussion

Knowledge concerning the movement and dispersal history of a species is fundamental to understanding modern patterns of diversification and speciation (Hickerson et al.

2010; Wiens & Donoghue 2004). Natural and anthropogenic processes such as geological landmass movements, climatic changes, sea level changes and commensal human movements facilitate the historical movement of many species including dogs

(Hewitt 2000; Hickerson et al. 2010; Kokko & López-Sepulcre 2006; Wiens 2004;

Wiens & Donoghue 2004). This study aims to investigate hypotheses concerning the origin, history and population structure of the dingo, Australia’s controversial apex terrestrial predator.

Phylogenetic analyses of the whole mtDNA and concatenated analyses of 13 nuclear genes provide strong evidence of at least two subdivided lineages of dingo in Australia.

Based on their biogeographic distribution we call these lineages the SE and NW lineages (Figure 2-1 and Figure 2-2). This is an important finding given the current persecution of the dingo in SE Australia. Further it suggests that management and conservation plans need to incorporate information concerning the current population structure of the dingo. Differing patterns of demography and selection between the two dingo lineages were observed in this study, further corroborating the presence of geographic subdivision (Table 2-3). Tajima’s D and other neutrality test results suggest that the NW lineage is experiencing population expansion whilst the SE lineage is experiencing population decline or contraction (Table 2-3). The SE lineage does appear to have a more restricted distribution and evidence of introgression from the NW

68 lineage. Plausibly, the NW lineage is, or has been, expanding into the SE lineage’s distribution. This may be particularly important if the lineages have unique ecologies.

Notably, we observed balancing selection within the NW lineage at two olfactory receptor genes, cfOR0007 and cfOR0034, suggesting these genes may play an important role in the divergence between the lineages and/or ongoing natural selection within this population (Table 2-3). Ecologically relevant factors plausibly impacting dingo lineage distribution patterns may include environmental gradients (Musiani et al.

2007; Rueness et al. 2003b; Stenseth et al. 2004) and neonatal dispersal (Sacks et al.

2004) or prey utilisation (Carmichael et al. 2001; Munoz-Fuentes et al. 2009; Musiani et al. 2007). In Australia, the establishment of a “dingo proof fence” in the southeastern corner of the continent may influence the movement of contemporary dingoes; however its erection between 1885 and 1950 seems likely too recent to be responsible for the geographical patterns observed in this study. More extensive biogeographic sampling, particularly in Northern Queensland, is needed to resolve the biogeographic distributions and possible introduction routes of the dingo lineages across Australia.

Evidence gathered during this study strongly suggests that inclusion of NGSD samples is necessary to fully investigate the evolutionary history of the Australian dingo. Whole mtDNA genome analyses suggest that the NGSD and the SE lineage of dingoes are distinct. In the mtDNA phylogeny posterior support for the node is low (Figure 2-1).

Nuclear phylogenies corroborate the mtDNA phylogeny indicating a close relationship between the NGSD and SE lineage of dingoes, with high levels of posterior support

(Figure 2-2). Topology testing provides strong evidence that dingoes are not monophyletic (Table 2-2), corroborating the mitochondrial and nuclear phylogenies

69 depicting a close relationship between the NGSD and SE dingo lineage (Figure 2-1 and

Figure 2-2). Increased sampling of NGSD from additional captive populations with a less constrained maternal background, or ideally wild and historical samples, may help further elucidate the relationships between dingo lineages and NGSD. Y chromosome data corroborates a close relationship between dingoes and NGSD (Ardalan et al. 2012;

Sacks et al. 2013). Archaeological and ethnographic evidence suggests that the NGSD once inhabited all of Papua New Guinea but were restricted to the mountains by the arrival of agriculture approximately 5,500 years BP (Koler-Matznick et al. 2004).

Morphologically the NGSD is similar to the dingo but can be differentiated on the basis of their smaller size and skull morphology (Koler-Matznick et al. 2004).

The close relationship of the SE lineage with the NGSD suggests the divergence between dingo lineages occurred before dispersal into modern Australia and/or outside

Australia, plausibly on Sahul (the landmass once incorporating Australia and Papua

New Guinea). There are two plausible hypotheses to explain the dingoes’ population structure in Australia, first that dingoes immigrated into Australia twice, and second that dingoes have undergone lineage sorting after an initial single introduction (of a population containing both lineages). The hypothesis that dingoes immigrated twice is supported by at least two lines of evidence. We observed that the SE and NW lineages are non-randomly distributed in the five populations sampled. FST and AMOVA results corroborate the mitochondrial and nuclear phylogenies indicating the presence of two geographically subdivided dingo lineages in Australia (Table 2-4 and Table 2-5). FST values also indicate strong geographic population differentiation particularly within the

SE lineage due to the geographic isolation of the Fraser Island population. AMOVA

70 analyses further corroborate the presence of strong population subdivision both at the lineage and population level in dingoes. The hypothesis of two dingo introductions is supported by Y chromosome data. Ardalan et al. (2012) observed the presence of two distinct paternal lineages, H3 and H60, with an East to West biased distribution.

Research on the Y chromosome by Sacks et al. (2013) also supports the idea of an East to West biased distribution of these same lineages. Sacks and colleagues found that the

H60 paternal lineage is more similar to Taiwanese dogs than to Island South East Asian dogs suggesting they split outside of Australia (Sacks et al. 2013). An alternative hypothesis is that the observed biogeographical pattern is the result of limited gene flow between populations and genetic drift rather than divergent immigration histories.

There are four discrepancies between the mtDNA and nuclear DNA phylogenies, involving dingoes that do not cluster consistently with geographic subdivision/patterns

(Figure 2-1 and Figure 2-2). The first is Alpine 5; this dingo was collected in the

Australian Alpine region but groups with the NW lineage for mtDNA and nuclear DNA phylogenies. Plausibly, Alpine 5 is a migrant into southeastern Australia. This is not unexpected as dingoes can be highly mobile (Corbett 1995; Fleming et al. 2001; Robley et al. 2010; Thomson et al. 1992). The second discordancy is Simpson 5 from the

Simpson Desert region; this dingo grouped with the NW lineage in the mtDNA phylogeny but clusters outside the SE/NGSD group in the concatenated nuclear DNA phylogeny. Potentially, Simpson 5 is the result of past male introgression from the SE lineage into the NW lineage. The third discrepancy is Fraser 3 and Fraser 4, both individuals collected from Fraser Island. These dingoes group with the SE lineage for mtDNA phylogenies but cluster within the NW lineage for nuclear DNA phylogenies.

71 Likely, Fraser 3 and Fraser 4 are hybrids between the two lineages resulting from a migration event from the NW lineage into the SE lineage. Potentially, Fraser 3 and

Fraser 4 are the result of past male dispersal events from the NW lineage into the SE lineage supporting the possibility that dingoes from the NW are expanding at the cost of

SE lineage animals. Indeed, differences in male and female dispersal rates or ranging distances have been observed in dingoes (Fleming et al. 2001; Robley et al. 2010;

Thomson et al. 1992) but it is not known if dingoes harbouring the different mtDNA types have any behavioural or ecological differences. The fourth and final discrepancy between the mtDNA and nuclear phylogenies are Alpine 3 and Alpine 4, whom cluster with the SE lineage for mtDNA phylogenies but group outside the main dingo phylogeny in multi-locus nuclear DNA analyses. This suggests that either Alpine 3 and

Alpine 4 are hybrids between dingoes and SE Asian dogs, or that Alpine 3 and Alpine 4 carry ancestral or conserved nuclear alleles that confound the nuclear phylogeny analysis. It is unlikely that Alpine 3 and Alpine 4 are hybrids between dingoes and SE

Asian dogs because both tested as pure dingoes (Wilton 2001; Wilton et al. 1999).

Next, we employed the whole mtDNA genome data to estimate the time the lineages diverged. We employed the published estimate of the whole mtDNA substitution rate of canids using the coyote-wolf fossil divergence time (Pang et al. 2009). If this substitution rate is correct, analyses of 25 dingoes from five populations suggest that the

SE dingo lineage and the NGSD mtDNA diverged in the ancient continent of Sahul more than 22,700 years BP. Indeed, the most conservative estimate of the tMRCA to the NW lineage is about 12,300 years and to the SE lineage is 9,700 years (Table 2-6).

These dates are well before to the Neolithic expansion, which occurred approximately

72 5,500 years BP (McEvoy et al. 2010; Milham & Thompson 1976; van Holst Pellekaan

2001, 2011). The divergence of these lineages is much older than expected but concordant with the mtDNA sequencing results on a 33,000 year old “Altai dog” skull, which suggested that dog domestication occurred earlier than previously hypothesised

(Druzhkova et al. 2013). One alternate explanation for these results is that the interspecific substitution rate derived from the coyote-wolf diverge is not appropriate for dating divergence within dingoes. We estimate that a two- to four-fold increase in the substitution rate would be required to reconcile with the fossil record if there were multiple introductions that diverged outside Australia. A two- to four-fold increase in mtDNA substitution rate is higher than expected in the dog (Pang et al. 2009), but a three- to ten-fold increase in mutation rate is observed in humans between phylogenetic and pedigree estimates of mutation rate (Endicott & Ho 2008; Langergraber et al.

2012).

We acknowledge mtDNA data provides a single estimate of the population history of these dingoes. As such, divergence times of the two dingo populations derived from mtDNA data should be treated with caution. Specifically, estimates may be reflective of the locus but not organismal divergence date (Ballard & Whitlock 2004). Four additional independent lines of evidence, however, suggest that dingoes immigrated into

Australia prior to the Neolithic expansion. First, Y chromosome data collected from dingoes and Island South East Asian dogs identified that dingoes are an older dog population that those in surrounding Island South East Asia, suggesting dingoes are the result of an earlier dog radiation, plausibly prior to the Neolithic expansion (Sacks et al.

2013). Second, despite uncertainty regarding the timing of dingo divergence, there is

73 increasing genetic evidence that suggests that dingoes and dogs evolved prior to the rise of agriculture. Freedman et al. (2014) found evidence that dingoes do not carry duplications of AMY2B, associated with starch metabolism that most other domestic dogs carry, further indicating that dingoes are unlikely to have been associated with agricultural cultures such as the Neolithic. Third, a recently discovered rock art painting of a dingo from Nawarla Gabarnmang in the Northern Territory, suggests they may have occurred in Australia earlier than previously thought (David et al. 2013).

Fourth, the Neolithic expansion into South East Asia, Papua New Guinea and Polynesia is characterised by human gene flow, introduction of agriculture and the presence of cultural items such as pigs, chickens and domestic dogs (Bocquet-Appel 2011;

Bramanti et al. 2009; Haak et al. 2010; Karafet et al. 2005; Larson et al. 2007;

Mulvaney & Kamminga 1999; Oskarsson et al. 2011; Sacks et al. 2013). However, cultural items associated with the Neolithic such as pigs and chickens were not brought to Australia until European colonisation in 1788 (Larson et al. 2007, 2010; Oskarsson et al. 2011). Additionally, there is no evidence of gene flow between Neolithic or East

Asian human populations and Indigenous Australians (McEvoy et al. 2010). Agriculture was also not present in Australia until the European colonisation in 1788 (Mulvaney &

Kamminga 1999). This ultimately suggests that the Neolithic expansion did not reach

Australia and was consequently not responsible for bringing dingoes to Australia. There is, however, one major issue with the hypothesis that dingoes colonised Australia before the Neolithic expansion: there are currently no dingoes in the fossil record prior to 3,500 years BP. One possible explanation for this result is that dingo fossils have not been well preserved in Northern Australia and Papua New Guinea. Certainly, there is a

74 general paucity of fossils in these regions due to soil acidity and tropical climatic conditions (Behrensmeyer 1978; Kidwell & Flessa 1996; Tappen 1994). We suggest it is unlikely that the mtDNA diverged long before the dingo colonised Australia because the lineages are clearly subdivided in Australia.

2.4.1 Conclusions

Mitochondrial and nuclear data presented in this study, and published in Y chromosome studies (Ardalan et al. 2012; Sacks et al. 2013) suggest the presence of at least two strongly subdivided lineages of dingo in Australia. This is of great importance to the conservation and management of the dingo in Australia. Populations representing both lineages should be protected and breeding programs should be designed incorporating this information. There is evidence that the dingo lineages are undergoing different demographic and selection pressures plausibly related to as yet unidentified differences in ecology. These data add to the growing body of evidence that suggests dingoes should be considered an indigenous Australian animal (Carthey & Banks 2012; David et al. 2013; Parsons & Blumstein 2010). Additionally the legislative status of the dingo may need to be updated from feral pest, and management practices may need to be adapted with an emphasis on conservation rather than exclusion and/or eradication. We hypothesise that dingoes immigrated to Australia at least twice. Our data suggest that plausibly dingoes and the NGSD diverged outside modern Australia and that the two lineages of dingo diverged before immigration into Australia. Genetic evidence, including molecular dating, suggests that potentially the human Neolithic expansion was not responsible for introducing the dingo and NGSD to Oceania.

75 Chapter 3

Biogeography of the native dog using uniparental markers: differences in female and male dispersal influences on biogeography

76 Abstract

Dingoes are a unique canid found only in Oceania. They play an important role in the ecological functioning of Australia as sole apex predator. They are currently threatened by hybridisation with domestic dogs in the wild. Genetic studies place the dingo within the domestic dog (Canis lupus familiaris) phylogeny. However, dingoes have been taxonomically reclassified as Canis dingo from Canis lupus familiaris (domestic dogs) on the basis of their divergent morphology, long-term reproductive isolation and ancient origins. Whole mitochondrial genome and nuclear data from five geographical populations showed evidence of at least two distinct lineages of dingo. Here, we present the data from a broader survey of dingoes around Australia using both mitochondrial and Y chromosome markers. Mitochondrial and Y chromosome data further support the presence of at least two geographically subdivided populations. The southeastern population is restricted to southeastern Australia whilst the northwestern population spreads from northern Queensland down to southern Western Australia. We present arguments that these two populations may be the result of two independent dingo introductions into Australia. There is also evidence of historical introgression from domestic dogs into dingoes, predominately within southeastern Australia.

77 3.1 Introduction

Dingoes are of worldwide interest due to their unique history of long-term isolation from modern domestic dogs (Canis lupus familiaris) and the likelihood that dingoes represent an ancient lineage of dog. This long-term isolation came to an end when

Europeans colonised Australia in 1788, bringing with them companion and working dogs. Since then, dingoes have been subject to hybridisation pressure from domestic dogs, particularly in regions where human populations are high. Their role as sole apex predator in Australia, however, subjects them to much controversy as the main threat to livestock.

Historically, the dingo is thought to have arrived on mainland Australia approximately

5,000 years before present (BP) (Gollan 1984; Macintosh 1975; Savolainen et al. 2004).

Although more recent molecular dating efforts based on mitochondrial divergence time suggest that it could have been in the proximity of 10,000 years BP (Oskarsson et al.

2011; Chapter 2). The minimum arrival time of dingoes is approximately 3,500 years

BP based upon the oldest fossil observed in southern Western Australia (Macintosh

1964). It is possible that dingoes and New Guinea Singing Dog (NGSD) came to

Oceania with humans. However, there is uncertainty concerning which human colonisation event they accompanied and whether dingoes were brought specifically to

Australia or immigrated via the prehistoric land bridge between Papua New Guinea and

Australia. It is commonly stated that dingoes followed humans to Australia as part of the Neolithic human expansion (Sacks et al. 2013; Savolainen et al. 2004), albeit a body of evidence contradicts this assumption. For example the lack of Neolithic cultural

78 items, such as pigs and chickens, in Australia prior to European colonisation (Larson et al. 2010; Oskarsson et al. 2011), lack of human genetic signatures indicating contact between South East Asia and Indigenous Australians (Haak et al. 2010; Karafet et al.

2005; McEvoy et al. 2010; van Holst Pellekaan 2001, 2011) and the finding that dingoes diverged from dogs before the agricultural era and only carry the two ancestral

Amylase gene copies (Freedman et al. 2014; Cairns unpublished data).

The species classification of dingoes is controversial due to the difficulty of resolving the relationships between dingoes, dogs and wolves using genetic methods. The dingo was classified as Canis antarticus (Kerr 1792) when first described, but was quickly renamed Canis dingo (Meyer 1793), and this binomial was confirmed by ICZN in 1957

(Crowther et al. 2014). By the 1990s the use of C. lupus familiaris or C. lupus dingo had become common. Most modern genetic studies suggest that dingoes fit within the domestic dog (C. lupus familiaris) phylogeny (Ardalan et al. 2012; Oskarsson et al.

2011; Savolainen et al. 2004; vonHoldt et al. 2010), and indicate a close relationship between dogs and wolves (Ardalan et al. 2011; Larson et al. 2012; Leonard et al. 2002;

Pang et al. 2009; Savolainen et al. 2002; Vilà et al. 1997, 1999; vonHoldt et al. 2010).

The Canis species are capable of hybridisation; this likely contributes to the difficulty in resolving the genetic relationships within the species complex given there is evidence of historical hybridisation and gene flow between the subspecies (Ardalan et al. 2011;

Hindrikson et al. 2012; Klütsch et al. 2011; vonHoldt et al. 2012). Additionally some authors present evidence that C. lupus is a species complex with cryptic species and distinct lineages (Aggarwal et al. 2007; vonHoldt et al. 2010, 2011), and that furthermore dogs, including dingoes and NGSD do not fit within the C. lupus lineage

79 (Freedman et al. 2014). Crowther et al. (2014) argued that the dingo should be re- confirmed as a separate species, C. dingo, rather than C. lupus familiaris or C. lupus dingo, given its divergent morphology, unique genetic characteristics, ancient origin, and long-term reproductive isolation. Logically, this is similar to the reasoning for the species classification of the NGSD (C. hallistromi) (Koler-Matznick et al. 2004).

Regardless of their classification and ethnography dingoes are of strong conservation value to present day Australia, so a thorough understanding of their biogeography is vitally important.

Dingoes are of strong conservation value for ecological, cultural and economic reasons.

Of particular importance to Australia’s ecological framework, the dingo is the sole remaining apex carnivore on the continental mainland and plays an important role as trophic regulator (Glen et al. 2007; Letnic & Koch 2010). Dingoes are theorised to play a role in the suppression of feral animal populations, including introduced species such as cats and foxes, via exclusion and direct predation (Brook et al. 2012; Johnson &

VanDerWal 2009; Letnic et al. 2009, 2010; Letnic & Dworjanyn 2011; Moseby et al.

2012). Through trophic cascade effects, dingoes may also play a protective role for native fauna (Letnic et al. 2009; Letnic & Dworjanyn 2011). Native marsupial species recognise dingo scents and undertake avoidance strategies indicating adaptation to the presence of the dingo, unlike introduced feral predators such as cats and foxes (Carthey

& Banks 2012). These ecological data suggest the dingo should be considered a native carnivore (Carthey & Banks 2012; Trigger et al. 2008).

80 The dingo is also closely linked to the Australian culture and its economy. It plays a significant role in the culture, theology, mythology and artwork of Indigenous

Australians (Smith & Litchfield 2009). The significant economic interest to the eco- tourism industry is most prominent on Fraser Island (Corbett 1995). Despite this the dingo, like many carnivores around the world, faces persecution particularly in agricultural pasturelands (Ripple et al. 2014).

In Australia, there are widespread management programs employing 1080 baiting and other lethal control measures (Fleming et al. 2001). Lethal control measures are shown to have variable success in reducing dingo activity and/or population sizes (Eldridge et al. 2002; Fleming et al. 2001). These control measures lead to pack destabilisation, increased hybridisation and possibly increased population sizes (Corbett 1988; Fleming et al. 2006; Glen 2010; Glen et al. 2007; Thomson 1992; Wallach et al. 2009, 2010).

Interestingly, increased livestock (cattle) losses have been reported following culling and baiting activities (Allen 2014). Overall, non-violent management strategies such as the use of livestock guardians and barrier fencing may actually prove more successful

(van Bommel & Johnson 2012). Understanding the underlying biogeography of the dingo may help governments and scientists to develop conservation programs that protect both genetic variation in the dingo and the interests of agriculture.

The most pressing concern to conservation and management programs is hybridisation between dingoes and modern domestic dogs. Methods of determining the purity of dingoes in the wild and captive populations have taken two routes: morphology or genetics. A study comparing the two methods indicated that morphology is a biased

81 indicator of purity, if skull measurements or genetic tests are not performed (Elledge et al. 2008). In the past, morphology has been used widely to classify dingoes as pure or hybrids, with indicators including skull measurements, body metrics and coat colour

(Corbett 1995, 2001; Newsome & Corbett 1982, 1985; Newsome et al. 1980).

Unfortunately, skull morphology is difficult to use in many circumstances, as to perform skull assessments the animal must be dead. Coat colour alone is not a reliable indicator of purity (Crowther et al. 2014; Elledge et al. 2008; Newsome & Corbett

1985). Morphology based assessments have identified that southeastern Australia has particularly high levels of hybridisation, likely due to the high density of human populations and their associated pet dogs and a prolonged history of baiting and culling

(Corbett 2001; Jones 2009; Newsome et al. 1980).

A genetic test developed and used since 1999 by Wilton and colleagues based on microsatellite allele frequencies, compares the microsatellite haplotypes of a sample to dog and dingo reference populations to estimate the dingo-domestic dog ancestry of an individual (Stephens 2011; Wilton 2001; Wilton et al. 1999). It is particularly useful for captive breeding and large-scale management programs, but could be improved by incorporating a larger dingo reference population, including geographical variation and historical samples. A large-scale genetic study on 4,500 dingoes from around Australia found that dingoes are genetically intact in most of regional Northern Territory and

Western Australia, but that dingoes from southeastern Australia have high levels of introgression (Stephens 2011). Mitochondrial DNA (mtDNA) control region studies suggest that the genetic testing method is robust, as few dingoes with non-dingo type mtDNA haplotypes have been identified (Oskarsson et al. 2011; Sacks et al. 2013;

82 Savolainen et al. 2004). European domestic dogs predominately carry H1 Y haplogroup types, which are also found in regions of South East Asia where European domestic dogs have been or are present (Brown et al. 2011; Ding et al. 2011; Natanaelsson et al.

2006; Sacks et al. 2013). Y chromosome studies have found limited evidence of domestic dog introgression in genetically verified dingoes (Ardalan et al. 2012; Sacks et al. 2013).

Studying the distribution of a species across geographic spaces or through time is commonly known as biogeography. Biogeographic knowledge regarding dingoes is limited. Skull morphology differences have also been noted between bioregions: alpine, desert and/or tropical (Corbett 1995). Morphological studies have also identified differences in the ratio of coat colours between different habitats (Corbett 1995;

Newsome & Corbett 1985). These observations might be a reflection of genetic subdivision, phenotypic plasticity or differences in nutrition. Conservation groups have long suggested that there are three varieties of dingo; alpine, desert and/or tropical

(Corbett 1995; Walters 1995). Genetic variation is of particular interest as it may elucidate cryptic patterns of variation and subdivision. A whole genome mitochondrial and nuclear DNA study identified the presence of at least two dingo lineages, southeastern (SE) and northwestern (NW), in Australia with a pattern of geographical subdivision (Chapter 2). Plausibly, these two populations are the result of multiple introductions into Australia. Previous genetic studies using mtDNA were unable to elucidate these patterns, as they were restricted to the mitochondrial control region

(Oskarsson et al. 2011; Savolainen et al. 2004). Two Y chromosome studies identified the presence of two paternal lineages in dingoes, H3 and H60. H3 is a Y haplogroup

83 widely observed in South East Asia whilst H60 is unique to dingoes and NGSD but closely related to H5 a haplogroup only observed in Taiwan (Ardalan et al. 2012; Sacks et al. 2013). However, uneven and restricted geographical sampling made biogeographically relevant conclusions difficult (Ardalan et al. 2012; Sacks et al. 2013).

One important concern with previous genetic studies on dingoes is that sampling designs were either restricted to certain geographical regions or too small, making it difficult to evaluate and interpret large-scale regional patterns. This study aims to investigate the biogeography of the dingo using mtDNA and Y chromosome markers in an effort to explore maternal and paternal evolutionary histories. We aim to investigate the distribution of the observed mitochondrial and Y chromosome haplogroups across

Australia. These data may shed light on migration and dispersal patterns in male and female dingoes on a continental scale.

3.2 Materials and methods

3.2.1 Canid sampling

In order to test hypotheses concerning biogeography, migration, male and female dispersal patterns and immigration routes, we sampled 127 dingoes broadly across

Australia and five NGSD from the North American captive population (Figure 3-1,

Appendix 2). Five of the dingoes were sampled from the captive dingo population. We also incorporated a dataset of Y chromosome and mitochondrial control region data from 191 male dogs, including 94 dingoes and 18 NGSD from Sacks et al. (2013).

84 10 0 −10 −20 −30

−40 Collected Samples canid sample

110 120 130 140 150

Figure 3-1 Map depicting geographic sampling of dingoes across Australia. Crosses represent individual samples. New Guinea Singing Dogs are depicted in Papua New

Guinea; however, samples were sourced from the North American captive population.

85 Tissue and/or blood samples were collected and all dingoes were pre-screened for genetic purity, using a microsatellite-based test for domestic dog introgression (Wilton

2001; Wilton et al. 1999). Only pure or genetically intact dingoes were assigned to this project (Stephens 2011; Wilton 2001; Wilton et al. 1999).

3.2.2 Mitochondrial gene analysis

Whole mtDNA and nuclear phylogenetic analyses found that there are at least two dingo lineages, with nine diagnostic mtDNA nucleotide differences between them (SE and NW, see Chapter 2). Two mtDNA regions harbouring diagnostic mutations and the mitochondrial control region were amplified and sequenced (Table 3-1). The two diagnostic regions were selected as they contained four of the nine differences between the SE and NW mitochondrial lineages. The regions were 676 bp (positions 7,685-8,361 including a region of ATP6 and ATP8) and 1,028 bp (positions 14,098-15,126 including a region of cytochrome b) in length. The mitochondrial control region is 582 bp (incorporating nucleotide positions 15,458-16,039 as in Savolainen et al. (2004)).

Table 3-1 PCR amplification primers and conditions for mitochondrial PCR amplification and sequencing of the dingo and NGSD Primer Name Sequence Nucleotide Reference position Diagnostic Region G8_F CCAATGATACTGAAGCTATG 7340 Designed by KMC Pair 1 G8_R ATTTTAGCAGGTTTGGTTAT 7915 Diagnostic Region G13_R CTAAAAGTCAGAATAGGCATT 15150 Designed by KMC Pair 2 P16_F TTCAGAACAATCGCACAACC 13973 Designed in Wilton Lab^ H15422 CTCTTGCTCCACCATCAGC 15422 Control region Savolainen et al. 2004 L16106 AAACTATATGTCCTGAAACC 16106 ^ Designed by M. Wong during his 2010 honours thesis (unpublished data) and supervised by AN Wilton.

DNA was extracted using a Qiagen DNeasy kit (Qiagen Sciences, Germantown, USA) and mitochondrial loci were amplified using PCR (Table 3-1). Briefly, PCR reactions were carried out in 25 uL volumes containing water, 5 X Crimson polymerase buffer

(New England Biolabs Inc., MA, USA), 1.5 mM of MgCl2, 6.25 pmol of each primer,

7.5 mM of dNTPs, 2.5 U of Taq DNA Polymerase (New England Biolabs Inc., MA,

USA) and 20-50 ng of DNA template. All PCR reactions were cycled using the following thermal profile: 98°C for 2 min, 95°C for 3 min (add Taq polymerase) then

95°C for 15 sec, 52°C for 1 min, 65°C for 1 min for 10 cycles, then 95°C for 15s, 52°C for 1 min, 65°C for 1 min (increase time by 5 sec each cycle) for 25 cycles followed by

65°C for 10 min.

PCR amplicons underwent ExoSAP-IT® (USB Amersham, Buckinghamshire, UK) purification prior to sequencing. Sanger sequencing was performed on the purified templates using BigDye terminator v3.1 (Applied Biosystems Inc., Foster City, USA).

Sequence chromatograms were visualised and aligned using Sequencher 5.1 (Gene

Codes corp., Ann Arbor, USA).

3.2.3 Y chromosome gene analysis

The iPLEX Sequenom MassARRAY system (Sequenom Inc., San Diego, USA) was used to genotype 29 single nucleotide polymorphisms (SNPs) from the non- recombining Y chromosome (NRY) region as described in Sacks et al. (2013). These 29

SNPs form a panel of markers enabling differentiation between most observed dog Y chromosome haplogroups (Ardalan et al. 2012; Brown et al. 2011; Ding et al. 2011;

88 Natanaelsson et al. 2006; Sacks et al. 2013). As in Sacks et al. (2013), we use H1 to refer to H1*, H2* and H1 haplotypes. Five dinucleotide-repeat single tandem repeats

(STR) were also genotyped from the NRY region: 650-79.2b, 650-79.3b, 990-35,

MS34CA, and MS41B as previously described (Brown et al. 2011; Sacks et al. 2013).

3.2.4 Biogeographic analyses

Median spanning networks were calculated in Networks v4.6 (Bandelt et al. 1999;

Forster et al. 2000) using the mitochondrial diagnostic region, mitochondrial control region and Y chromosome datasets. As in Sacks et al. (2013), the Median Joining (MJ) algorithm with default settings was used (r=2, ε=0). Mitochondrial networks were created for the concatenated diagnostic region and control region separately. Control region data were analysed separately to allow incorporation of and comparison to the existing dingo control region dataset (Savolainen et al. 2004; Oskarsson et al. 2011;

Sacks et al. 2013). Since the control region is not phylogenetically informative in dingoes it was not included the mitochondrial diagnostic region analysis (Chapter 2). Y chromosome networks were calculated using concatenated SNP and STR data. Y chromosome SNPs and STRs were weighted as described by Sacks et al (2013).

Briefly, STRs were weighted as: 650-79.2b = 5, 650-79.3b = 2, 990-35 = 9, MS34CA =

6, MS41B = 1 and SNP loci = 90 (Brown et al 2011; Sacks et al 2013). Networks were drawn using our collected data and an additional dataset including 112 Oceanic samples from Sacks et al. (2013).

To further investigate the relationship between the dingo and NGSD (identified in

Chapter 2), we ran Bayesian analyses in Beast v1.7.4 (Drummond et al. 2012), allowing

89 us to estimate the posterior probability value of the inferred relationship. Chapter 2 found that the posterior probability value was low (0.4) suggesting uncertainty regarding the position of the NGSD lineage within dingoes. The Bayesian analysis was conducted on a set of 124 dingoes plus 5 NGSD; three dingoes were excluded due to

PCR amplification difficulties. Bayesian analyses were run in Beast v1.7.4 (Drummond et al. 2012) under a constant coalescent model with a strict clock (Chapter 2; Brown &

Yang 2011). All runs were optimised for MCMC chain steps to ensure the estimated sampling size was above 200 in Tracer 1.5 (Rambaut & Drummond 2007). Sampling occurred every 5,000 steps with a 10% burn-in. The resulting maximum clade credibility tree was midpoint rooted.

The biogeographic distribution of each individual belonging to each mitochondrial or Y chromosome haplogroup was then plotted onto maps using the maps package

(Brownrigg et al. 2014) in R (R Development Core Team 2010), allowing visualisation of the distribution of the mitochondrial and Y chromosome lineages across Australia.

Simple contingency table analyses were used to evaluate whether the distribution of Y chromosome haplogroups between the mitochondrial lineages was non-random.

To investigate the relationship of Y chromosome haplotypes found in dingoes and

NGSD with those found in Island South East Asia, a network was calculated based upon data from 173 dingoes, 20 NGSD and 79 South East Asian dogs, incorporating our dingo and NGSD dataset as well as the dataset from Sacks et al. (2013). The resulting network was colour coded relative to geographical region.

90 3.2.5 Neutrality tests

To investigate whether genetic variation present within the mitochondrial genome departs from the expectations of neutrality Tajima’s D, Fu and Li’s F* and Fu and Li’s

D* statistics were calculated in DnaSP v 5.10.1 (Fu 1997; Fu & Li 1993; Tajima 1989;

Librado & Rozas 2009). These statistics can be used to investigate the presence of demographic or selective pressures acting upon the molecular evolution of a DNA sequence. Significantly negative values indicate the presence of purifying selection and significantly positive values suggest balancing selection. Non-significant negative values indicate population expansion whilst positive values indicate population decline.

Values close to zero indicate neutral evolution, i.e. no indication of demographic or selective pressures. These neutrality statistics were calculated for all dingoes and then specific dingo populations separately.

3.3 Results

3.3.1 Biogeographic analyses

When ignoring indels we observed 12 mitochondrial control region (CR) haplotypes with 3 novel CR haplotypes in 124 dingoes (three dingoes were excluded due to PCR difficulties) and 5 NGSD (Table 3-2). The novel haplotypes (din31, din32 and din33) were found in 1-4 individuals and differed by 1-2 nucleotide substitutions from A29.

One dingo carried the A9 haplotype thought to have arisen in dingoes independently from dogs (Savolainen et al. 2004). A single dingo out of 124 carried A17, a CR haplotype hypothesised to be introgressed from domestic dogs (Savolainen et al. 2004).

91 Incorporating all the CR data from previously published studies to our own yielded a star-shaped genetic network (Figure 3-2).

A total of 39 mitochondrial diagnostic region haplotypes were observed in 124 dingoes and 5 NGSD (Table 3-2). None were consistent with non-dingo mitochondrial lineages.

The 5 NGSD all carried the same haplotype. The mitochondrial diagnostic region network displays a more interesting pattern consistent with the presence of two mitochondrial haplogroups, SE and NW, in Australia (Figure 3-3). The genetic network also corroborated the close relationship between the SE lineage and NGSD. The biogeographic distribution of the two mtDNA haplogroups across Australia was plotted, indicating strong geographic subdivision with limited mixing between the two populations (Figure 3-4). The SE mtDNA lineage was restricted to Fraser Island and the southeastern coastal region of Australia (Queensland, New South Wales, Australian

Capital Territory and Victoria), while the NW mtDNA lineage was widespread from

Western Australia to northern/central Queensland and south into South Australia. A single NW mtDNA lineage individual was observed within the Australian Alpine region. Within the captive dingo population both NW and SE haplogroups were observed; captive animals were not plotted on the map.

Table 3-2 Mitochondrial and Y chromosome haplotype data for 124 dingoes and 5 NGSD CR haplotype CR haplotype MtDNA Dingo Name Sex State (collapsed) (gaps) lineage MtDNA type Y-chr type Alpine 1 M VIC A209 A209 SE a9 H60-k11 Alpine 2 M NSW A29 A29 SE a3 H1-k1 Alpine 3 M ACT A29 A29 SE a3 H1-6t Alpine 4 F VIC A29 A179 SE a3 Alpine 5 M VIC A29 A179 NW d23 H60- k11 Fraser 1 F QLD A29 A179 SE f1 Fraser 2 M QLD A29 A179 SE f1 H60- n25 Fraser 3 F QLD A29 A179 SE f1 Fraser 4 M QLD A29 A179 SE f1 H60- n25 Fraser 5 F QLD A29 A179 SE f1 Fraser 6 F QLD A29 A179 SE a1 Gibson 1 F WA A203 A203 NW d5 Gibson 2 F WA A29 A29 NW d5 Gibson 3 M WA A200 A200 NW d5 H60- n24 Gibson 4 F WA A29 A29 NW d5 Gibson 5 F WA A200 A200 NW d5 Kimberley 1 F WA A29 A29 NW d5 Kimberley 2 M WA A9 A9 NW d7 H3- n4 Kimberley 3 M WA A9 A9 NW d7 H3-n20 Kimberley 4 F WA din27 din27 NW d5 Kimberley 5 M WA A29 A29 NW d5 H3- n4 Simpson 1 F SA A29 A179 NW d5

Simpson 2 M NT A200 din35 NW d5 H60 -9k

Table 3-2 Mitochondrial and Y chromosome haplotype data for 124 dingoes and 5 NGSD CR haplotype CR haplotype MtDNA Dingo Name Sex State (collapsed) (gaps) lineage MtDNA type Y-chr type Simpson 3 F SA A200 A200 NW d4 Simpson 4 F SA A200 A200 NW d14 Simpson 5 F NT A200 A200 NW d5 Simpson 6 M NT A29 A29 NW d3 H60- n25 Captive 1 M Captive A29 A29 SE a3 H1-k7 Captive 2 M Captive A29 A205 NW d5 H60-k8 Captive 3 M Captive A29 A29 SE a3 H1-n7 Captive 4 M Captive A29 A29 SE a3 H3-12d Captive 5 M Captive A29 A198 NW d5 H60-n24 Central Australia 1 M QLD din32 din32 NW d20 H1-6q Central Australia 10 M NT A29 A29 - H60-n29 Central Australia 11 M NT A17 A17 - H3-6z Central Australia 12 M SA din32 din32 NW d21 H60-k10 Central Australia 13 M SA din31 din31 NW d5 H60-n25 Central Australia 14 M SA A29 A179 NW d22 H1-6q Central Australia 4 M SA din32 din32 NW d5 H1-6q Central Australia 2 M QLD din32 din32 NW d20 H1-6q Central Australia 3 M NT A29 A179 NW d5 H1-6q Central Australia 5 F NT A29 A29 NW d1 Central Australia 6 F NT - - NW d2 Central Australia 7 M NT A200 A200 NW d9 H60 -0i Central Australia 8 M NT A200 A200 NW d9 H60-n24 Central Australia 9 M NT A29 A29 NW d5 H60-n25 Dubbo M NSW A29 A199 NW d8 H3-k9 Inglewood F QLD - - NW d16

Table 3-2 Mitochondrial and Y chromosome haplotype data for 124 dingoes and 5 NGSD CR haplotype CR haplotype MtDNA Dingo Name Sex State (collapsed) (gaps) lineage MtDNA type Y-chr type Moree M NSW A29 A208 SE a10 H1-6t NGSD 1 F PNG A79 A79 NGSD ng1 NGSD 2 F PNG A79 A79 NGSD ng1 NGSD 3 F PNG A79 A79 NGSD ng1 NGSD 4 M PNG A79 A79 NGSD ng1 H60- k10 NGSD 5 M PNG A79 A79 NGSD ng1 H60-k10 Northeastern 1 F QLD A29 A199 NW d16 Northeastern 10 M QLD A29 A199 NW d5 H60 -9k Northeastern 11 M QLD A29 A199 NW d5 H60-n27 Northeastern 2 M QLD A29 A207 NW d16 H60-n28 Northeastern 3 F QLD A201 din34 NW d16 Northeastern 4 M QLD A29 A207 NW d16 H60- n27 Northeastern 5 F QLD A29 A179 NW d16 Northeastern 6 F QLD A29 A179 NW d16 Northeastern 7 F QLD A29 A179 NW d16 Northeastern 8 M QLD A29 A179 NW d16 H60- n24 Northwestern 1 M WA A200 din36 NW d5 H60-n17 Northwestern 10 F WA A29 A29 NW d12 Northwestern 11 M WA A29 A29 NW d5 H3- n4 Northwestern 12 F WA A29 A29 NW d5 Northwestern 13 F WA din27 din27 NW d18 Northwestern 14 F WA A29 A29 NW d5 Northwestern 15 F WA A200 A200 NW d5 Northwestern 16 F WA A29 A29 NW d24

95 Northwestern 17 F WA A29 A29 NW d5

Table 3-2 Mitochondrial and Y chromosome haplotype data for 124 dingoes and 5 NGSD CR haplotype CR haplotype MtDNA Dingo Name Sex State (collapsed) (gaps) lineage MtDNA type Y-chr type Northwestern 18 M WA A29 A29 NW d10 H3-n21 Northwestern 19 M WA A202 A202 NW d11 H60-k11 Northwestern 2 M WA A29 A29 NW d15 H60-n22 Northwestern 20 M WA A210 din33 NW d5 H60-n25 Northwestern 21 M WA A203 A203 NW d5 H60-n22 Northwestern 22 M WA A29 A29 NW d5 H60-n25 Northwestern 23 M WA A29 A29 NW d5 H60-k3 Northwestern 3 M WA A29 A29 NW d5 H3-12d Northwestern 4 M WA A29 A29 NW d19 H3-n21 Northwestern 5 F WA A29 A29 NW d10 Northwestern 6 M WA A202 A202 NW d6 H3- n21 Northwestern 7 F WA A29 A179 NW d13 Northwestern 8 F WA A200 A200 NW d5 Northwestern 9 F WA A29 A29 NW d15 Southeastern 1 M QLD A213 A213 SE a10 H1- 1c Southeastern 10 F NSW A29 A29 SE a2 Southeastern 11 F NSW A29 A29 SE a2 Southeastern 12 F NSW A29 A29 SE a2 Southeastern 13 F NSW A29 A29 SE a11 Southeastern 14 F ACT A29 A29 SE a3 Southeastern 15 F VIC A29 A29 SE a3 Southeastern 17 F QLD A29 A179 SE a1 Southeastern 18 F VIC A29 A29 SE a2 Southeastern 19 F VIC A29 A29 SE a2

96 Southeastern 2 F QLD A29 A179 SE a10

Table 3-2 Mitochondrial and Y chromosome haplotype data for 124 dingoes and 5 NGSD CR haplotype CR haplotype MtDNA Dingo Name Sex State (collapsed) (gaps) lineage MtDNA type Y-chr type Southeastern 20 M ACT A29 A29 SE a3 H1-k1 Southeastern 21 M ACT A29 A29 SE a3 H1-6t Southeastern 22 M ACT A29 A29 SE a3 H1-k1 Southeastern 23 M ACT A29 A29 SE a3 H1-k1 Southeastern 24 M ACT A29 A29 SE a3 H1-k1 Southeastern 25 M ACT A29 A29 SE a3 H1-6t Southeastern 26 M QLD A29 A29 SE a7 H60-9k Southeastern 27 M NSW A29 A29 SE a3 H1-6t Southeastern 28 M NSW A29 A29 SE a3 H1-6t Southeastern 29 M NSW A29 A179 SE a3 H60-n24 Southeastern 3 F NSW A29 A29 SE a8 Southeastern 30 M NSW A29 A29 SE a3 H3- k2 Southeastern 31 M NSW A29 A29 SE a1 H3-k9 Southeastern 32 M NSW A29 A29 SE a3 H1-k1 Southeastern 33 M NSW A29 A29 SE a1 H3-k9 Southeastern 34 M NSW A29 A29 SE a1 H3-k9 Southeastern 35 M NSW A29 A29 SE a6 H1-6t Southeastern 36 M NSW A29 A29 SE a1 H60-n25 Southeastern 37 M NSW A29 A29 SE a1 H1-k6 Southeastern 38 M NSW A29 A208 SE a5 H3-k9 Southeastern 39 M NSW A29 A29 SE a1 H60-k4 Southeastern 4 M ACT A29 A29 SE a2 H1-6t Southeastern 40 M NSW A29 A29 SE a1 H1-6t Southeastern 41 M NSW A29 A29 SE a1 H1-k6 Southeastern 42 M NSW A29 A29 SE a3 H21*-7d 97

A79 din33 A202 A201

A17 A209 A9 A211 din23

A213

A206 A29 A203 A204

din27

A212 din30 din32

A200

din31 din22

Figure 3-2 Median spanning network based on mitochondrial control region data from

450 dingoes and 23 NGSD. The network was calculated including our dataset as well as

94 dingoes and 18 NGSD from Sacks et al. (2013) and 232 dingoes and 3 NGSD from

Oskarrson et al. (2011). Red colour indicates NGSD samples whilst blue indicates a haplotype hypothesised to be introgressed from domestic dogs. Circles are proportional to the number of individuals carrying that haplotype.

Figure 3-3 Median spanning network based upon the mitochondrial diagnostic region (1706 bp) in 124 dingoes and 5 NGSD. Black

colouration indicates NW lineage haplotypes, orange SE lineage haplotypes, red NGSD haplotypes and purple indicates captive

individuals. Strokes across branches indicate the presence of SNP mutations fixed between the mtDNA lineages, branch lengths are relative

to the number of mutations separating mtDNA haplotypes. 100

●

● −10

●

● ● ●● ●● −15 ●● ● ● ● ●● ● ● ● ●

●● ● ● ● −20 ● ● ● ● ● ●●

● ● ● ● ● ●● −25 ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● −30 ● ● ●● ●● ● ● ● ● ● ●

● ●● −35 ●●●●●● ● ● ● ●●● ● ●●●●● −40 Mitochondrial Clades ● SE Clade ● NW Clade −45 120 130 140 150

Figure 3-4 Biogeographical map of 119 dingoes and their mitochondrial lineage designation. Black circles indicate NW lineage haplotypes and orange SE lineage haplotypes. Captive dingoes (five) are not included on the map.

101 To further investigate the relationship between the dingo and NGSD, a Bayesian analysis was conducted on the combined sample of 129 animals. This included 124 dingoes and 5 NGSD (Table 3-2). This analysis corroborated the whole mtDNA genome Bayesian phylogenetic analyses suggesting that the NGSD is more closely related to the SE dingo lineage than the NW lineage (Chapter 2), with an increased posterior probability node support of 0.84 (Figure 3-5).

We observed 30 Y chromosome haplotypes in our dataset of 79 dingoes and 2 NGSD

(Table 3-2). Y chromosome network analysis identified three main haplogroups present within dingoes and NGSD: H1, H3 and H60 (Figure 3-6). A contingency table analysis, with two columns (mitochondrial lineage) and three rows (Y chromosome haplogroup), suggests that the distribution of Y chromosome haplogroups between the mitochondrial lineages was non-random in dingoes (Χ2=18.1, d.f.= 2, p<0.001). To further investigate the distribution of the Y chromosome haplogroups across Australia, we incorporated the dingo and NGSD data from Sacks et al. (2013) resulting in a total of 194 samples (173 dingoes and 20 NGSD; Figure 3-7). We observed an additional 6 Y chromosome SNP-

STR haplotypes (Figure 3-7). The 20 NGSD sampled between the two datasets all carried H60 haplotypes. When Y chromosome haplogroup information was plotted on a map (Figure 3-8), we observed that H1 was largely restricted to the southeastern region of Australia, H3 was restricted to the southeastern and Kimberley regions, and H60 was predominantly found throughout northern, western and central Australia. H1 was predominately found in southeastern Australia. Of the four H3 haplogroup alleles observed in the Kimberley region, all were endemic except H3_12d, which was also observed in southeastern Australia.

102 Southeastern Captive Alpine

Southeastern

Alpine Southeastern 0.98 Captive

Southeastern 0.98

1 Alpine 0.93 Southeastern

Fraser Southeastern 0.97 Fraser

Southeastern

0.84 Moree 1 NGSD

1 Central 0.85 Northeastern 0.95 Inglewood 0.93 Northeastern Northwestern Central Northwestern Dubbo Northwestern Central Northwestern Central Kimberley Northeastern Kimberley Northwestern Simpson Northwestern Gibson 0.98 Central Gibson Northwestern Central Gibson Captive Simpson Northwestern Simpson Central Gibson 0.98 Alpine 5 Northwestern Northwestern Simpson Northwestern Central 0.98 Captive 0.93 Kimberley Northwestern Central 3.0E-4 Figure 3-5 Bayesian analysis of 124 dingo and 5 NGSD mtDNA diagnostic region

(1706 bp) sequences. Analyses constructed using a GTR + G + I substitution model and a coalescent model with constant population size in BEAST v1.7.4 (Drummond et al.

2012). Posterior probability values are reported below nodes and values less than 0.6 are not shown. Colours represent geographical sampling population, black for NW, orange for SE and red for NGSD. The scale bar indicates an estimate of the average number of substitutions per site between two nodes.

103 H60 n27 n24 H3

n25

H21 H1

Figure 3-6 Median spanning network based Y chromosome SNP and STR haplotypes for 79 dingoes and 2 NGSD. Black colouration

indicates NW mtDNA lineage individuals, orange SE mtDNA lineage individuals, red NGSD individuals and white indicates unknown

mtDNA lineage. Strokes across branches indicate the presence of Y chromosome SNP mutations differentiating between Y chromosome

haplogroups. Branch lengths are relative to the number of STR mutations between Y-chromosome haplotypes. 104

H60

H21

Figure 3-7 Median spanning network based Y chromosome SNP and STR haplotypes for 173 dingoes and 20 NGSD. Black colouration

indicates dingoes from this study, red NGSD individuals and white indicates dingo samples from Sacks et al. (2013). Strokes across

branches indicate the presence of Y chromosome SNP mutations differentiating between Y chromosome haplogroups. Branch lengths are

relative to the number of STR mutations between Y-chromosome haplotypes. 105

To investigate the relationship between Y chromosome haplotypes observed in dingoes,

NGSD and South East Asian dogs, a MJ network was calculated based upon a combined dataset of 272 samples. This comprised of 173 dingoes, 79 from our dataset and 94 from Sacks et al. 2013; 20 NGSD, 2 from our dataset and 18 from Sacks et al.

2013 and 79 South East Asian dogs from Sacks et al. (2013) (Figure 3-9). We observed that the H1 and H3 haplotypes found in dingoes were largely unique (not shared with

South East Asian dogs). Further investigation, however, suggests that H1-1c, H1-6t,

H1-n7 and H1-6q were observed in European domestic dog breeds (Brown et al. 2011;

Sacks et al. 2013). Three novel H1 haplotypes were observed (H1-k1, H1-k7 and H1- k5) that were unique to dingoes.

3.3.2 Neutrality tests

Tajima’s D statistics were calculated for all dingoes as grouped by mitochondrial lineage using the mtDNA diagnostic region (Table 3-3). Statistics could not be calculated for the NGSD as all individuals carried the same mtDNA sequence. The NW lineage statistics were found to be significantly negative, indicating the presence of purifying selection and/or population expansion. Statistics calculated for the SE lineage were negative but not significant. This suggests that the different mitochondrial lineages may be experiencing different demographic/selective pressures.

106 −10 −15 −20 −25 −30 −35

−40 Y−Chromosome Haplotype H1 H3 H60 H21 −45 120 130 140 150 Figure 3-8 Biogeographical map of 168 dingoes and their Y chromosome haplogroup designation. Colouration indicates Y chromosome haplogroup; red for H60, blue for

H3, purple for H21 and green for H1. Only wild dingoes were plotted on the map.

107 H60

H21

H5 H3 H1

H12

H14 H13 H6

Figure 3-9 Median spanning network based Y chromosome SNP and STR haplotypes for 173 dingoes, 20 NGSD and 79 South East Asian

dogs. Colouration indicates geographical region/species type: black for dingoes, red for NGSD individuals, green for Taiwan, gray for

Thailand, blue for Brunei, yellow for Philippines and purple for Bali. Connection lengths are proportional to the number of mutations

(either STR or SNP). Strokes across branches indicate the presence of Y chromosome SNP mutations differentiating between Y

108 chromosome haplogroups. Branch lengths are relative to the number of STR mutations between Y-chromosome haplotypes.

Table 3-3 Nucleotide variation and Neutrality statistics on mitochondrial DNA (1706 bp) from 124 dingoes. π θ Hd Tajima’s D Fu and Li’s F* Fu and Li’s D* NW Lineage 8.9 x 10-4 3.70 x 10-3 0.77 -2.43** -4.58 ** -4.52 ** SE Lineage 1.08 x 10-3 1.40 x 10-3 0.79 -0.65 -1.91 -2.10 (**) indicates p <0.02

109

109 3.4 Discussion

Mitochondrial and Y chromosome data collected in this study support the hypothesis that dingoes have an Asian origin, as the haplotypes observed in dingoes are most closely related to those found in Asian dogs (Ardalan et al. 2012; Brown et al. 2011;

Oskarsson et al. 2011; Sacks et al. 2013; Savolainen et al. 2004; Chapter 2). This is consistent with previous mitochondrial, nuclear and Y chromosome studies which observed a close relationship between dingoes, NGSD and Asian dogs (Ardalan et al.

2012; Brown et al. 2011; Oskarsson et al. 2011; Pang et al. 2009; Sacks et al. 2013;

Savolainen et al. 2004; vonHoldt et al. 2010). We found no evidence of an Indian origin for dingoes as suggested by some authors (Pugach et al. 2013; Pugach & Stoneking

2013). Dingoes harbour mitochondrial haplotypes which are more closely related to those observed in South East Asia that those in India (Pang et al. 2009). Instead, our data suggest the likely introduction route of the dingo and NGSD into Oceania is through South East Asia with an Asian origin.

As has been suggested in Chapter 2, biogeographic mitochondrial data corroborate the presence of at least two matrilineal dingo populations: SE and NW (Figure 3-3).

Additionally, the geographical distribution of these two mitochondrial lineages exhibited strong geographical subdivision (Figure 3-4). Only a single instance of discordance between mitochondrial lineage and geographic origin was observed, indicating that migration between the mitochondrial lineages is limited. This suggests that matrilineal migration between the populations may be biased west to east. Previous mitochondrial studies did not observe the presence of population structure in dingoes

110 due to restricted sampling of the mitochondrial control region (Oskarsson et al. 2011;

Sacks et al. 2013; Savolainen et al. 2004).

The geographical distribution of the three Y chromosome haplogroups H1, H3 and H60, is strikingly similar to that of the mitochondrial lineages, supporting the presence of at least two geographically subdivided populations (Figure 3-4, Figure 3-8 and Chapter 2).

The distribution of Y chromosome haplotypes between the two mitochondrial lineages is non-random, strongly corroborating the presence of at least two geographically subdivided populations as first observed in Chapter 2. Introgression between the NW and SE populations again appears to be largely west to east biased, with few H3 haplogroup individuals found in northern, western or central Australia (excluding the

Kimberley region which carries unique H3 haplotypes). The Kimberley region harbours

H3 haplotypes which are different to those observed in southeastern Australia suggesting it is likely they are the result of shared ancestry but not recent introgression.

Conversely, there are some individuals in southeastern Australia harbouring H60 haplogroup types, likely the result of male dispersal from the NW population into the

SE population. Previous Y chromosome studies observed similar distribution patterns; however, uneven geographical sampling prevented them from observing the large-scale subdivision present in our dataset (Ardalan et al. 2011; Sacks et al. 2013).

Combined, the mtDNA and Y chromosome data indicate the presence of at least two discrete populations of dingo, the NW (H60/H3) and the SE (H3/H1). The lack of intermediate haplotypes between the two matrilineal lineages despite large-scale geographical sampling suggests that the described populations structure is due to

111 historical events, i.e. multiple introductions rather than isolation by distance.

Additionally, the Y haplogroups H3 and H60 are not immediately related (Figure 3-9;

Ardalan et al. 2012; Brown et al. 2011; Natanaelsson et al. 2006; Sacks et al. 2013).

The H3 haplogroup is observed in South East Asia; however, of the eight haplotypes observed in dingoes, all were endemic except one, indicating shared ancestry with a history of isolation between dingoes and South East Asian dogs (Figure 3-9; Brown et al. 2011; Sacks et al. 2013). The H60 haplogroup, on the other hand, is unique to dingoes and NGSD (Figure 3-9; Ardalan et al. 2012; Brown et al. 2011; Sacks et al.

2013). It is most closely related to H5, found in Taiwan (Figure 3-9). This suggests that the two paternal haplogroups found in dingoes have different origins and are likely the result of multiple introductions.

The restricted distribution of the SE matrilineal lineage may provide further clues into the colonisation history of the dingo. As hypothesised in Chapter 2, plausibly there were two separate introductions of dingoes, with the first spreading down into southeastern

Australia, and the second (at a later date) expanding and swamping out the population

(closely related to the SE population) in northern Australia. Mitochondrial divergence estimates are consistent with two dingo introductions, possibly at different times

(Chapter 2). An alternative hypothesis is that the SE population is the result of an introduction event into southeastern Australia. A single introduction, as suggested by

Savolainen et al. (2004) is unlikely given the strong biogeographical subdivision at both maternal and paternal markers and the divergent evolutionary relationships between the two populations.

112 Biogeographic patterns within Australia provide insight into the modern dispersal and migration of dingoes. Matrilineal migration is observed to be rare (Figure 3-4); however, paternal introgression seems to be more widespread (Figure 3-8). Fraser

Island dingoes appear to share NW paternal lineage ancestry but carry SE mtDNA types. These patterns are likely a factor of male dispersal; male dingoes and dogs range more widely and are more likely to disperse to new areas (Pal et al. 1998; Thomson et al. 1992). Human-mediated dispersal may also be a factor in facilitating the movement of dingoes, by breaking apart pack structures through culling/baiting management practices (Corbett 1988; Fleming et al. 2006; Glen et al. 2007; Thomson 1992; Wallach et al. 2009, 2010). The divergent origins of the different matrilineal and paternal lineages suggest that modern barriers such as the dingo fence are not responsible for the historical pattern of geographical subdivision.

Ongoing geographic subdivision despite no obvious physical barriers suggests that there may be some ecological or behavioural factors at play. In coyotes neonatal habitat has been attributed to patterns of subdivision (Sacks et al. 2004, 2008). In wolves differences in prey specialisation are hypothesised to be the reason for cryptic population subdivision (Carmichael et al. 2001; Munoz-Fuentes et al. 2009; Musiani et al. 2007; Weckworth et al. 2011). It is possible that the dingo fence plays a role in limiting modern dispersal between populations. Plausibly, the west to east biased dispersal pattern observed here is the result of NW population dingoes moving into vacant niches opened up by extensive culling and baiting practices in southeastern

Australia (Stephens 2011). Neutrality test results suggest that the two dingo populations

113 may be experiencing different demographic and/or selective pressures (Table 3-3 and

Chapter 2).

In Chapter 2, Whole mtDNA genome analyses suggested that the NGSD and SE lineage of dingoes were distinct; however, support for the node was low. In this study we observed the same close relationship between the NGSD and SE mtDNA lineage of dingoes with a much higher level of support (Figure 3-5). Mitochondrial network analyses also corroborated the close relationship between the SE lineage and NGSD

(Figure 3-3). Y chromosome data, however, suggest that NGSD share a closer paternal relationship with the NW lineage than the SE lineage (Figure 3-6 and Figure 3-7). As identified by Sacks et al. (2013) NGSDs all harboured a H60 Y chromosome haplogroup, which is consistent with a small captive founding population. It is possible that the captive NGSD population is not representative of the full genetic variation of wild NGSD. Given that the captive NGSD population in Northern America was founded by ~1-7 male progenitors this seems likely (Koler-Matznick et al. 2004). These data may suggest the revision of the NGSD’s taxonomic name to Canis dingo hallistromi, although the exact nature of the relationship between the NGSD and the two dingo populations is still uncertain, given conflicting maternal and paternal histories. Further research into the relationship between the NGSD and two dingo populations using autosomal markers, uniparental markers and wild NGSD samples, may help further elucidate the historical biogeography and evolution of the dingo and

NGSD.

114 The H1 Y chromosome haplogroup is considered to be a European domestic dog haplogroup (Ardalan et al. 2012; Brown et al. 2011; Ding et al. 2011; Sacks et al.

2013), and it is often observed in domestic dog breeds or South East Asian dogs thought to have breed ancestry. The presence of H1 within dingoes suggests the occurrence of paternal introgression from European domestic dogs into dingoes. An alternative hypothesis is that the H1 haplogroup is ancestral to dingoes. Three of the seven H1 STR haplotypes were unique and so far only observed in dingoes. The other four H1 STR haplotypes were shared with breed dogs (Sacks et al. 2013), providing strong evidence of European dog introgression. The presence of the H1 haplogroup in genetically tested

“pure” dingoes suggests that either H1 is ancestral (pre-European colonisation) or the introgression is historical (post-European colonisation) rather than modern. It is unlikely that the introgression is modern because the genetic test is capable of detecting hybridisation events on a recent time-scale (Cairns et al. 2011; Wilton 2001; Wilton et al. 1999). The uniparental inheritance of the Y chromosome means that a single hybridisation event will be reflected in the paternal lineage of a dingo despite extensive backcrossing. The lack of non-dingo-like mitochondrial lineages suggests that introgression from domestic dogs is predominately male mediated/biased.

The distribution of the H1 haplogroup in southeastern Australia further suggests it is likely the result of introgression (Figure 3-8). First, domestic dogs have been present in southeastern Australia for a longer period of time, having arrived with European colonists in 1788, allowing for a longer period of sympatry with dingoes (Corbett

1995). Second, southeastern Australia has the densest human population and areas with dense human populations are generally associated with higher incidences of

115 hybridisation (Stephens 2011). Thirdly, lethal management strategies such as baiting and culling are widespread in southeastern Australia due to the sheep industry (Fleming et al. 2001). Fatal management strategies are believed to lead to increased levels of hybridisation due to the breakdown of pack structure (Corbett 1988; Fleming et al.

2006; Wallach et al. 2009).

The identification of likely historical paternal introgression from domestic dogs into dingoes is troubling for the conservation of the species. Some may argue that dingoes with any level of introgression are not worth conserving, despite the important ecological role they fulfil. Certainly, genetic integrity should be an important factor in choosing individuals or populations for conservation; however, perhaps not the most important factor. A controversial study by Garrick et al. (2012) suggests that backcrossing hybrids may allow for the reconstruction of an extinct or highly threatened species, in that case an extinct Galápagos tortoise. They argued that introgressed individuals with intact nuclear and mtDNA genetic signatures are of high conservation value to targeted breeding efforts. In other species facing introgression from domestic forms such as the American Bison, trade-offs between maintaining populations free of introgression, healthy ecological functioning and preserving the full range of genetic diversity are being debated (Freese et al. 2007; Halbert 2003; Halbert & Derr 2007) and

Wildcats (Hertwig et al. 2009). Perhaps in wild dingo populations expectations of 100% genetic integrity are not realistic, particularly in southeastern Australia, and management strategies should be employed to preserve and improve genetic integrity along with the existing geographic variation within the dingo. This has been described as a “hybrid reduction” strategy where hybrids of detectable (or a certain introgression

116 level) are removed (Halbert 2003). It is important to note, however, that this strategy will not re-create pure dingoes, but rather reduce the level of detectable hybrids.

Patterns of genetic subdivision in other large carnivores have been linked to ecologically relevant characteristics such as neonatal dispersal (Sacks et al. 2008; Sacks et al. 2004), prey specialisation (Carmichael et al. 2001), environmental climes

(Rueness et al. 2003a, 2003b; Stenseth et al. 2004) and sociality (Randall et al. 2010). It is plausible that the two dingo populations have differing ecological characteristics that could be relevant to the conservation and management of the species. Of utmost importance, these data suggest that conservation and management efforts should be focused on maintaining the existing population structure and thus treat the two populations as distinct conservation units. The presence of the H1 haplogroup in southeastern Australia has important implications for conservation and future management strategies. Namely, it highlights the importance of inhibiting further hybridisation. Neutering male dogs and/or restricting them from reproducing with wild dingoes may help achieve this. The southeastern population of dingoes is under particular extinction pressure from both fatal management strategies and hybridisation; steps should be taken to preserve this population before it is too late. Currently state and federal legislation do not protect the dingo sufficiently and allows widespread fatal control measures in southeastern Australia (Davis 2001; Downward & Bromell 1990;

Fleming et al. 2001, 2006). Revision of legislation must be achieved to reflect the ecological, cultural and taxonomic importance of the dingo; balancing the need to conserve this enigmatic canine with any agricultural concerns.

117 Two significant knowledge gaps that remain are whether there is phenotypic variation between the two dingo populations and whether the two dingo populations were distributed differently in the past. Conservation groups have long described the presence of multiple morphological varieties of dingo generally alpine, desert and/or tropical; however, it is not clear whether this phenotypic variation is associated with the genetic subdivisions or phenotypic plasticity. A future study of morphological and phenotypic variation as well as genetic variation may help uncover this question. Of additional interest is whether the distribution of the two dingo populations has changed significantly, which may be elucidated with fossil, sub-fossil and historical sampling.

These data may also help elucidate the immigration and/or colonisation history of dingoes. The unique genetic identity of dingoes from the Kimberley region is also interesting, given they carry a range of H3 Y haplotypes not observed elsewhere in

Australia. Further sampling of dingoes from the Kimberley may be particularly important for future studies concerning the biogeography and origin of dingoes.

The mitochondrial control region identifies limited population structure within the dingo (Figure 3-2), in contrast to the population structure observed in other regions of the mtDNA (Figure 3-3, Figure 3-5 and Chapter 2). This data suggests that the mitochondrial control region is not useful for investigating population structure and variation within the dingo or NGSD. Further it provides only limited utility in comparisons between dingoes and domestic dogs, as it does not reflect the multi-lineage history of the dingo. Future studies should endeavour to incorporate the mitochondrial diagnostic region, described here, at the very least.

118 3.4.1 Conclusions

This study corroborates the presence of at least two dingo populations in Australia. It is plausible, given the divergent evolutionary histories of these populations that they are the result of two independent introductions into Australia. These two populations are characterised by differing mitochondrial, nuclear and Y chromosome signatures. The two dingo populations are geographically subdivided, with one restricted to the southeast of Australia and the other widespread across central, northern and western

Australia. There is evidence of historical, post-European colonisation, paternal introgression from domestic dogs into the SE dingo population. Conservation, management and legislative practices need to be revised to reflect the presence of two dingo populations and to limit future hybridisation.

119 Chapter 4

Elucidating biogeographical patterns in Australian native canids using whole

genome genotyping technologies

120 Abstract

Dingoes play a strong role in Australia’s ecological framework as the apex predator but are under threat from hybridisation and agricultural control programs. Government legislation lists the conservation of the dingo as an important aim, yet little is known about the biogeography of this enigmatic canine, making conservation difficult.

Mitochondrial and Y chromosome DNA studies observe the presence of population structure within the dingo. Here, we present the data from Illumina HD Canine chip genotyping for 23 dingoes from five regional populations, and 5 New Guinea Singing

Dogs in an effort to explore patterns of biogeography using whole genome data. Whole genome single nucleotide polymorphism data observes the presence of 3 distinct dingo populations subject to geographical subdivision: southeastern, Fraser Island and northwestern. The New Guinea Singing Dog is observed to be divergent from dingoes but closely related. Inbreeding coefficients identified that the Fraser Island dingo population is highly inbred. For the Fraser Island population, a known reservoir of pure and genetically distinct dingoes this is extremely troubling. Fraser Island dingoes may be genetically compromised and in need of rescue; current lethal management strategies that do not consider genetic information should be suspended until further data can be gathered. New Guinea Singing Dogs in captivity are also under threat of high inbreeding, not unexpected due to the small foundation population. The confirmation of different geographically subdivided dingo populations using maternal, paternal and autosomal markers suggests the need for these populations to be treated as discrete conservation units.

121 4.1 Introduction

Dingoes are a controversial topic in Australia, treated as either a feral pest or a native animal (Fleming et al. 2001). Like many other top-order carnivores, dingoes pose a risk to human livestock activities and are thus extensively managed in the livestock grazing regions of southeastern Australia (Ripple et al. 2014). On the other hand, dingoes are considered a native species and protected in national parks (Davis 2001; Downward &

Bromell 1990; Elledge et al. 2006; Fleming et al. 2001). Legislation differs between

Australian states and territories causing a discontinuity in management and conservation practices across the continent (Davis 2001; Downward & Bromell 1990; Fleming et al.

2001; Trigger et al. 2008). Little is known about the biogeography and genetic identity of the dingo, with few published scientific studies. Most scientific research focuses upon the ecology of the dingo. Genetic evidence to date suggests dingoes are at severe risk of extinction through hybridisation with European domestic dogs (Elledge et al.

2008; Stephens 2011; Wilton 2001; Wilton et al. 1999). Currently, dingoes are listed as vulnerable on the IUCN Red List (Corbett 2008). Conservation and management programs, adequately informed by scientific knowledge, must be developed to protect the genetic identity of the dingo before it is lost.

Current genetic phylogenies based upon mitochondrial, Y chromosome and whole genome single nucleotide polymorphism (SNP) data suggest that dogs, wolves and dingoes form a monophyletic group (Ardalan et al. 2012; Brown et al. 2011; Oskarsson et al. 2011; Pang et al. 2009; Sacks et al. 2013; Savolainen et al. 2004; vonHoldt et al.

2010a). Evidence from copy number variation at the Amylase locus indicates that

122 perhaps dogs and dingoes are not direct descendants of grey wolves but rather descended from some other common ancestor (Freedman et al. 2014). Australian scientists have argued that due to the history of long isolation, divergent genetics, unique ecology and morphology, the dingo should be considered a distinct species,

Canis dingo, like the New Guinea Singing Dog (NGSD), Canis hallistromi (Crowther et al. 2014). Scientists agree that the dingo is an ancient dog-like canid, despite controversy regarding the exact taxonomy. The dingoes’ value as an ancient breed in understanding the evolution, domestication and movement of dogs cannot be overstated.

The dingo is thought to have arrived in Australia at least 5,000 years before present

(BP). The earliest fossil evidence was found in Western Australia and has been dated, through radiocarbon dating of the strata, to 3,500 years old (Macintosh 1964, 1975).

Other anthropological and fossil evidence suggests the minimum arrival time of the dingo is approximately 4,000-5,000 years BP (Gollan 1984; Macintosh 1975; Milham

& Thompson 1976; Mulvaney 1975). Genetic dating estimates suggest that the dingo may have diverged from their ancestor either 5,000 years BP (Sacks et al. 2013;

Savolainen et al. 2004), 4,600-18,300 years ago (Oskarsson et al. 2011), or at least

10,000 years BP (Chapter 2).

Domestic dogs are thought to have diverged from their wolf (or wolf-like) ancestor approximately 16,000 years ago (Pang et al. 2009). Fossil and genetic evidence suggests that plausibly dogs diverged from wolves (or a wolf-like canid) 18,800 -32,100 years ago (Druzhkova et al. 2013; Thalmann et al. 2013). Amylase copy number variation analyses suggest that dogs were likely domesticated before the rise of

123 agriculture (Freedman et al. 2014). Plausibly dingoes, which don’t have increased amylase copies, split from domestic dogs prior to domestication (Freedman et al. 2014).

Particularly high copy numbers of amylase in domestic dogs suggest strong selection pressure and differentiation occurred during the rise of agriculture (Axelsson et al.

2013). Genetic comparisons between dingoes and other canids suggest that dingoes have an affinity with Asian wolves and dogs, hinting that their heritage is most likely

Asian (vonHoldt et al. 2010a). It is hypothesised that dingoes migrated into Australia through South East Asia (Ardalan et al. 2012; Oskarsson et al. 2011; Sacks et al. 2013).

Dingoes play a strong role in shaping the ecosystems of Australia as a tropic regulator

(Fleming et al. 2001; Glen & Dickman 2005; Glen et al. 2007; Letnic & Koch 2010).

Dingoes are the only remaining top-level predator on the continental mainland, following the extinction of the Thylacine (Johnson & Wroe 2003). They exert top-down control on large herbivores such as kangaroos and emu (Caughley et al. 1980; Dickman

1996; Dickman et al. 2009; Fillios et al. 2010; Johnson et al. 2007; Letnic & Crowther

2012; Letnic et al. 2012; Marsack & Campbell 1990; Pople et al. 2000; Robertshaw &

Harden 1986; Shepherd 1981; Wallach et al. 2010) and indirectly protect native small- medium bodyweight marsupials (Johnson et al. 2007; Letnic et al. 2009a, 2009b;

Wallach et al. 2009a).

Dingoes may also play a role in reducing the impact of introduced feral mesopredator pests such as foxes and cats by exclusion and direct predation (Cupples et al. 2011;

Johnson & VanDerWal 2009; Kennedy et al. 2012; Letnic & Dworjanyn 2011; Letnic et al. 2010; Letnic & Koch 2010; Mitchell & Banks 2005; Moseby et al. 2012; Wang &

124 Fisher 2012). Notwithstanding strong scientific evidence, the mesopredator release hypothesis and positive effect of dingoes is controversial with extensive ongoing scientific debate (Allen 2010; Allen et al. 2011a, 2011b; Allen & Fleming 2012;

Fleming et al. 2012a, 2012b; Glen 2012; Johnson & Ritchie 2012; Letnic et al. 2011).

Despite evidence of the important role dingoes play in a natural context, they are heavily controlled by lethal measures such as 1080 baiting and shooting in an effort to protect human agricultural activities (Fleming et al. 2001, 2006). Dingoes in southeastern Australia, the “sheep grazing” belt south of the dingo fence, are most strongly controlled (Fleming et al. 2001). However, lethal control and management practices may not be having the desired effect. There is evidence of increased predation of small-medium weight prey and reduced mesopredator control, caused by the disintegration of dingo packs and increasing solitary dingo numbers (Brook et al. 2012;

Thomson 1992). There is also evidence that lethal management practices may always not decrease dingo population sizes and as a result of pack destabilisation may cause increased levels of hybridisation (Wallach et al. 2009b). There is some evidence that cattle graziers may experience heavier stock losses when dingoes are culled/baited, probably owing to the collapse of social structures (Allen 2014). On Fraser Island, dingoes are controlled to limit negative human-dingo interactions (DEHP 2013).

Management decisions regarding dingoes are often made without scientific, particularly genetic, data regarding the purity, identity and genetic health of a population.

Knowledge concerning the genetic health of natural populations, particularly isolated or threatened populations, is important for the development of effective conservation and

125 management programs. Specifically inbreeding may influence the vulnerability of a natural population to environmental or demographic fluctuations (Keller & Waller

2002). Inbreeding coefficients are calculated by comparing the observed homozygosity to the expected homozygosity of genetic markers (Frankham & Ralls 1998; Keller &

Waller 2002). Severe inbreeding is believed to be a particular risk to conservation efforts because it may lead to negative fitness consequences through the accumulation of fixed recessive or partially recessive deleterious mutations (Frankham & Ralls 1998).

Inbreeding depression, a possible result of severe inbreeding, can lead to the extinction of a population (Ballou et al. 1998; Frankham & Ralls 1998; Keller & Waller 2002;

Laikre & Ryman 1991; O’Grady et al. 2006; Trinkel et al. 2011; Xu et al. 2007). An increase in inbreeding within domestic dogs has been associated with the increased incidences of genetic disease (Leroy et al 2009; Ólafsdóttir & Kristjánsson 2008). Wild canid populations and outbred domestic dog breeds have inbreeding levels closer to zero (Fain et al. 2010; Huson et al. 2010; Kennedy et al. 1991; Stenglein et al. 2011).

There are particular concerns regarding the genetic viability of the Fraser Island dingo population given their low effective population size and conservation significance

(Appleby & Jones 2011; Stephens 2011; Woodall et al. 1996).

Genetic studies in the past have focused largely upon uniparental markers such as mitochondrial and Y chromosome markers. In the past, mitochondrial control region studies asserted that modern dingo populations were likely the result of a single very homogeneous foundation population, possibly even a single pregnant female

(Savolainen et al. 2004). However, Sacks et al. (2013) and Ardalan et al. (2012) observed the presence of two divergent paternal genetic lineages within dingoes. Large-

126 scale biogeographical patterns could not be discerned by these studies due to uneven geographical sampling. In Chapter 2 and Chapter 3, the presence of geographical subdivision between populations was observed using both mitochondrial and Y chromosome markers. Chapter 2 and Chapter 3 argued that the presence of multiple divergent lineages of dingo, which are geographically subdivided, might indicate multiple introductions. However, uniparental markers may be maternally or paternally biased and show different evolutionary patterns to autosomal markers (Galtier et al.

2009; Humphries & Winker 2011; Rubinoff & Holland 2005; Toews & Brelsford 2012;

Zink & Barrowclough 2008).

In fact the evolution of uniparental markers can be divergent to that of the “species”. In the past utilisation of single locus autosomal markers would be used to corroborate evolutionary histories depicted by mitochondrial or Y chromosome markers. Chapter 2 used thirteen autosomal fragments to corroborate the mitochondrial phylogeny, observing a similar phylogeny between mitochondrial and autosomal markers.

However, there is emerging evidence that gene trees formed from a small set of coding regions may not be representative of the whole “organism” evolutionary history either

(Brito & Edwards 2009; Degnan & Rosenberg 2009; Rokas et al. 2003). Whole genome

SNP assays and next-generation sequencing technologies allow scientists to capture genetic variation across the entire genome visualising a more complete picture of the organism or individual’s evolutionary signature.

Here we present the first whole genome SNP array dataset for a set of dingoes from five geographically distinct populations and NGSD. We aim to further investigate

127 previously observed patterns of geographic subdivision in the dingo and NGSD. We wish to examine patterns of genetic diversity and variation within the dingo populations to help inform conservation and management strategies. Additionally, we aim to investigate and identify ancestry informative markers for future biogeographical surveys.

4.2 Methods and Materials

4.2.1 Canid sampling

To investigate patterns of biogeography, we sampled 25 wild dingoes from five geographical populations: The Kimberley (Western Australia), The Gibson Desert

(Western Australia), The Simpson Desert (Northern Territory), Fraser Island

(Queensland) and the Australian Alpine region (Australian Capital Territory, Victoria and New South Wales) (Appendix 2). Five NGSD from a captive population in North

America were also included given the close relationship between dingoes and NGSD

(Chapter 2 and Chapter 3).

DNA was extracted from blood or tissue samples using the Qiagen DNeasy Blood and

Tissue kit according to the manufacturer’s instructions (Qiagen Sciences, Germantown,

USA).

128 4.2.2 Illumina HD Canine genotyping

Samples were genotyped on the 170K Illumina HD Canine SNP array (Illumina Inc.,

San Diego, USA) at the Cornell Genomics Core Facility (Cornell University, Ithaca,

USA). The 170K Illumina HD Canine SNP array contains SNPs identified through comparisons between the dog reference genome and shotgun sequence data from nine other domestic dog breeds, four wolves and a coyote (Lindblad-Toh et al. 2005). The

SNP array was designed to provide coverage across the entire genome with approximately 70 SNPs per Mb (Illumina Inc., San Diego, USA). Genotypes were called using GenomeStudio (Illumina Inc., San Diego, USA). Quality control and filtering was conducted in PLINK v1.7 (Purcell et al. 2007). Specifically, individuals missing more than 10% of SNPs were excluded, SNP sites with more than 10% missing were excluded and SNPs with a minor allele frequency of less than 5% were excluded.

A sex check analysis was performed in PLINK v1.7 (Purcell et al. 2007) to investigate inbreeding as estimated by homozygosity at the X chromosome and confirm gender assignments.

For clustering and phylogenetic analyses, SNP data from the X and Y chromosomes were excluded because differences in effective population size, recombination rates and inheritance patterns may bias clustering and phylogenetic analyses.

4.2.3 Inbreeding and homozygosity

Inbreeding coefficients (FIS), examining differences in observed and expected

129 homozygosity levels for each individual were calculated in PLINK v1.7 (Purcell et al.

2007) using the filtered SNP dataset.

4.2.4 Clustering analysis

Admixture software v1.23 (Alexander & Lange 2011) was used to perform maximum likelihood clustering analyses. Admixture was chosen as it is more capable of dealing with large SNP datasets than other clustering programs such as Structure and FRAPPE

(Alexander & Lange 2011; Alexander et al. 2009). These programs work to evaluate K, the estimated number of population clusters present within a dataset and Q the population identity/proportion of each individual (Alexander & Lange 2011; Alexander et al. 2009). Clustering analyses were run with the following conditions: 10-fold cross- validation and iterations for each K were run until the change in the log likelihood value was below 0.1. Ten independent runs of each K value were completed, each using different random seeds. Comparing the cross validation errors for each K value and between the independent runs, the best K value was chosen (Alexander & Lange 2011).

Clumpp v1.1.2 was used to compare the Q matrices of the ten independent runs for the best K (Jakobsson & Rosenberg 2007). An average Q-plot for the best K was constructed using Distruct v1.1 (Rosenberg 2004). A map depicting the population assignment of each sample was created using the maps package (Brownrigg et al. 2014) in R (R Development Core Team 2010).

4.2.5 Principal components analysis

A principal components analysis (PCA) was performed on the filtered SNP dataset in

130 PLINK v1.9 (Purcell & Chang 2014; Purcell et al. 2007). The top 20 eigenvalues and eigenvectors were calculated. The percentage variation that each principal component

(PC) vector accounts for was calculated using the following formula: eigenvalue/(Σ of all eigenvalues) × 100. The top three PCA eigenvectors, accounting for the largest percentage variance, were plotted against each other using the rgl package (Adler et al.

2014) in R (R Development Core Team 2010).

4.2.6 Genetic distances

Classical multidimensional scaling (MDS) for 3 factors was also performed using

Identical By State (IBS) allele matrices calculated in PLINK v1. 7 (Purcell et al. 2007) and plotted in R (R Development Core Team 2010) using the rgl package (Adler et al.

2014). A Neighbour Joining tree was constructed in MEGA 6 (Tamura et al. 2013) from the IBS allele matrices (1-IBS) to investigate genetic distances between individuals. FST values between the four population clusters were calculated in Admixture v1.23

(Alexander & Lange 2011).

4.2.7 Phylogenetic analyses

The phylogenetic relationships between the 23 dingoes and 5 NGSD were further investigated using SNPhylo (Lee et al. 2014) a pipeline for constructing maximum likelihood (ML) trees from genome wide SNP datasets. The SNP dataset was pruned for invariant SNPs and the remaining ancestry informative SNPs were concatenated together in SNPhylo (Lee et al. 2014). Concatenation is a method of combining

131 sequences from multiple genetic loci and is particularly useful for intraspecific datasets where divergences may be recent (Gadagkar 2005). However concatenation may introduce biases as a result of rate heterogeneity, differences in gene tree topology and/or recombination (Edwards 2009). As such phylogenies should be treated conservatively.

A ML tree using a Hidden Markov Model was constructed by DNAml (Felsenstein

2005) as implemented in the SNPhylo pipeline (Lee et al. 2014). Non-parametric bootstrapping with 6000 repetitions was performed on the ML tree using Phangorn

(Schliep 2011). Ancestry informative markers, as identified by SNPhylo (Lee et al.

2014), were further explored using PLINK v1.7 (Purcell et al. 2007) to investigate their utility in future genetic studies. In particular, allele frequencies and association statistics were calculated. SNPs with Wald test values of p <10-4 were considered to be strongly associated with population structure in the dingo and NGSD.

To further investigate the relationship between dingoes and NGSD a second dataset was formed including 12 wolf (Canis lupus) samples from Axelsson et al. (2013; specifically Wlf_LUb3, Wlf_22800, Wlf_LU1655, Wlf_Lub1, Wlf_19785,

Wlf_LU1656, Wlf_22802, Wlf_22803, Wlf_22810, Wlf_LU1657, Wlf_Lub2 and

Wlf_22809). These wolf samples were used as outgroup taxa. The dataset was merged in PLINKv1.7 (Purcell et al. 2007) then SNPhylo (Lee et al. 2014) was used to create a pruned and ancestry informative concatenated sequence for each sample. ML analyses were run in raxmlGUI (Silvestro & Michalak 2012) first without topological constraints

(H0) and second with SE dingoes and NGSD forced to be monophyletic (HA) as

132 observed in Chapter 2. All analyses in raxmlGUI (Silvestro & Michalak 2012) were run with a GTR + G substitution model and 2000 bootstrap replicates. Per site log likelihood values were calculated for the best unconstrained and constrained trees in raxmlGUI (Silvestro & Michalak 2012). CONSEL (Shimodaira and Hasegawa 2001a) was then used to calculate the Kishino-Hasegawa (KH) test (Kishino and Hasegawa

1989), Shimodaira-Hasegawa (SH) test (Shimodaira and Hasegawa 1999) and

Approximately unbiased (AU) test (Shimodaira and Hasegawa 2001b) values between the unconstrained and constrained trees. Significant KH, SH or AU test p values indicate that the HA can be rejected.

4.3 Results

4.3.1 Illumina HD Canine genotyping

Of the 25 dingoes genotyped, two were excluded from the analyses, one for failing to adequately run and one for having more than 10% missing SNPs. This left the study with a total of 23 dingoes from 5 geographical populations (Table 4-1). Five NGSD samples were genotyped as well and incorporated in this analysis (Table 4-1). The total genotyping rate before filtering was 0.974169, and after filtering was 0.99017. The total number of SNPS before filtering was 173,662, and after filtering was 58,512.

A sexcheck performed in PLINK v1.7 (Purcell et al. 2007) identified that many of the female dingoes had high F values (F>0.2). High F values indicate excessive homozygosity at the X chromosome, either due to inbreeding or an artefact of the

133 dingoes’ divergent evolutionary history (Table 4-1).

4.3.2 Inbreeding and homozygosity

Inbreeding statistics (FIS) calculated based on 58,512 SNP loci suggest that some dingo populations are inbred (Table 4-1). Particularly striking is the finding that the four

Fraser Island dingoes had very high FIS values of 0.647-0.732 indicating an extreme level of inbreeding. Similarly the NGSD population was highly inbred with FIS values ranging from 0.457-0.671. The Alpine and Simpson Desert populations had the lowest average FIS statistics.

4.3.3 Clustering analyses

Admixture analyses on 58,512 SNP loci found that the K value with the lowest cross- validation error across ten independent runs was K=4, suggesting the presence of four population clusters in the SNP dataset. (Figure 4-1) The Q-plot depicts the population identity of each canid individual (Figure 4-2). There is some evidence of admixture between population clusters in some individuals, particularly Alpine 1, Alpine 5 and the three Simpson Desert dingoes. FST values between the four population clusters indicate a high level of differentiation and low gene flow between the populations (Table 4-2).

Samples were plotted on a map to depict the geographical pattern of subdivision and distribution of population clusters (Figure 4-3).

134 Table 4-1 Inbreeding (FIS) and X chromosome homozygosity (F) for 23 dingoes and 5 NGSD calculated from Illumina SNP data from 58,512 sites. Geographical ID population Sex F FIS Average FIS

Alpine 1 Alpine M 0.90 0.002 Alpine 2 Alpine M 0.89 0.038 Alpine 3 Alpine M 0.94 0.137 Alpine 4 Alpine F 0.77 0.331 0.089 Alpine 5 Alpine M 0.90 -0.060 Fraser 3 Fraser M 0.91 0.732 Fraser 4 Fraser M 0.86 0.647 Fraser 5 Fraser F 0.74 0.720 0.700 Fraser 7 Fraser M 0.88 0.702 Gibson 1 Gibson F 0.77 0.439 Gibson 2 Gibson F 0.31 0.122 Gibson 3 Gibson M 0.92 0.100 Gibson 4 Gibson F 0.21 0.250 Gibson 5 Gibson F 0.35 0.188 0.240 Northwestern 2 Gibson M 0.93 0.339 Kimberley 1 Kimberley F 0.29 0.271 Kimberley 2 Kimberley M 0.90 0.218 Kimberley 3 Kimberley M 0.89 0.142 Kimberley 4 Kimberley F 0.24 0.270 0.215 Northwestern 9 Kimberley F 0.11 0.175 Simpson 1 Simpson F 0.09 0.161 Simpson 2 Simpson M 0.94 0.097 0.138 Simpson 5 Simpson F 0.20 0.157 NGSD A NGSD M 0.94 0.457 NGSD B NGSD M 0.92 0.671 NGSD C NGSD M 0.93 0.594 NGSD D NGSD F 0.38 0.620 0.561 NGSD E NGSD M 0.93 0.461 Note: two dingoes of the original 25 were excluded from this study one for failing to adequately run (Fraser 6) and one for having more than 10% missing SNPs (Simpson 3).

135

0.9

0.8

0.7 Cross Validation Error

0.6 0 1 2 3 4 5 6 7 K value

Figure 4-1 Cross-validation errors for each K-value averaged across ten independent runs in Admixture v1.23 (Alexander & Lange 2011). Error bars represent standard error calculated across the ten runs. Cross-validation error was lowest for K=4.

136 NGSD Southeastern Fraser Island Northwestern

NGSD A NGSD C

NGSD B NGSD E Alpine 5

NGSD D Alpine 3 Alpine 2 Alpine 1

Alpine 4 Fraser 3 Fraser 5 Gibson 5 Simpson 2

Fraser 4 Simpson 3 Northwestern 2 Gibson 4 Fraser 7 Simpson 1 Kimberley 4 Northwestern 9 Kimberley 1 Kimberley 3 Gibson 2

Kimberley 2 Gibson 1 Gibson 3

Figure 4-2 Maximum likelihood population clustering analysis on 23 dingoes and 5

NGSD at 58,512 SNP loci. Average Q-plot for K=4 constructed in Distruct v1.1

(Rosenberg 2004). Each column represents an individual and the proportion population cluster identity. Population clusters are represented by colours: green for New Guinea

Singing Dog, red for southeastern, purple for Fraser Island and blue for northwestern.

137

Table 4-2 FST values between dingo and NGSD populations. Calculated by Admixture (v1.23) based on 58,512 SNP loci. NGSD Fraser Island Southeastern NGSD - - - Fraser Island 0.408 - - Southeastern 0.354 0.61 - Northwestern 0.238 0.431 0.421

138

−10 −20 −30

Population Clusters NGSD

−40 Southeastern Fraser Island Northwestern 100 110 120 130 140 150

Figure 4-3 Geographical map depicting sampling location of each sample and majority population cluster identity. NGSD samples (plotted in Papua New Guinea) are from a captive North American population.

139 4.3.4 Principal components analysis

PCA found that the top three PC vectors account for 21.3% (eigenvalue = 5.343),

15.05% (eigenvalue = 3.774) and 9.84% (eigenvalue = 2.468) of the genetic variance respectively. Plots of the top three PC vectors indicate the presence of four population clusters: SE, FI, NW and NGSD (Figure 4-4-1). Interestingly, the FI population seems to cluster with the NW dingo population (Figure 4-4-2). Additionally the SE population appears to be more heterogeneous than the other populations.

4.3.5 Genetic distances

The MDS plot displays a similar clustering pattern as the PCA and clustering analyses, corroborating the presence of four population clusters in the dataset (Figure 4-5). A

Neighbour Joining tree constructed in MEGA 6 to visualise genetic distances between the dingo and NGSD samples, based upon IBS matrices (1-IBS), identified the presence of four population clusters (Figure 4-6). As with the PCA, FI dingoes appear to share a closer relationship with the NW dingoes than SE dingoes. Also, dingoes form a monophyletic group with NGSD grouped outside.

140 Dingo Population Northwestern Southeastern Fraser Island NGSD

0.4

0.2 0.6 PCA 3 0.4 0.0 0.2 PCA 2 0.0 −0.2 −0.2 −0.2 0.0 0.2 0.4 PCA 1

Figure 4-4-1 Principal Components Analysis (PCA) based upon filtered whole genome

SNP genotypes (58,512 sites) for 23 dingoes and 5 NGSD in 3-dimensions. Shown here are the top three PC eigenvectors. PCA performed in PLINK v1.9 (Purcell & Chang

2014; Purcell et al. 2007) and plot drawn using the rgl package (Adler et al. 2014) in R

(R Development Core Team 2010).

141 −0.1 0.1 0.3 0.5 0.3 PCA 1 21.30% 0.1 −0.1 0.5 0.3 PCA 2

0.1 15.05% −0.1 0.4

PCA 3 0.2 9.84% 0.0

−0.1 0.1 0.3 −0.2 0.0 0.2 0.4 −0.2

Figure 4-4-2 Principal Components Analysis (PCA) based upon filtered whole genome

SNP genotypes (58,512 sites) for 23 dingoes and 5 NGSD. This figure serves to illustrate the variance and pattern contained within each PC vector using 2-dimensions.

Shown here are the top three PC eigenvectors. PCA preformed in PLINK v1. (Purcell &

Chang 2014; Purcell et al. 2007) and plot drawn using the rgl package (Adler et al.

2014) in R (R Development Core Team 2010). Colours represent population clusters: red for SE dingoes, purple for FI dingoes, blue for NW dingoes and green for NGSD.

142 Dingo Population Southeastern Northwestern Fraser Island NGSD 0.15

0.20 MDS 3

0.05 0.10

0.00 MDS 2 −0.10 −0.10−0.05 0.00 0.10 0.20 MDS 1

Figure 4-5 Classical MDS analysis with three factors based upon filtered whole genome SNP genotypes (58,512 sites) for 23 dingoes and 5 NGSD. Analysis was performed in PLINK v1.7 (Purcell et al. 2007) and plot drawn using the rgl package

(Adler et al. 2014) in R (R Development Core Team 2010).

143 Gibson 1 Gibson 5 Gisbon 4 Northwestern 2 Gibson 2 Gibson 3 Northwestern 9 Kimberley 3 Kimberley 4 Kimberley 2 Kimberley 1 Simpson 2 Simpson 5 Simpson 1 Fraser 3 Fraser 7 Fraser 5 Fraser 4 Alpine 1 Alpine 4 Alpine 5 Alpine 3 Alpine 2 NGSD A NGSD D NGSD E NGSD B NGSD C

0.05

Figure 4-6 Neighbour Joining tree depicting genetic distances (1-IBS) between 23 dingoes and 5 NGSD based on 58,512 SNP loci. Tree constructed in MEGA6 (Tamura et al. 2013) based upon 1-IBS matrix calculated in PLINK v1.7 (Purcell et al. 2007) depicts genetic distances between individuals and populations based upon filtered whole genome SNP genotypes from 23 dingoes and 5 NGSD. Colours represent population clusters: red for SE dingoes, purple for FI dingoes, blue for NW dingoes and green for

NGSD.

144 4.3.6 Phylogenetic analyses

SNPhylo identified 4,913 SNPs that were variable and ancestry informative. Of these a total of 460 SNPs were identified has having significant Wald test values (p<10-4) indicating a strong association between these SNPs and geographical population. These

SNPs may be useful for developing a SNP assay to investigate population structure in a larger geographic survey of dingoes.

The unrooted ML tree constructed in DNAml (Felsenstein 2005) identified four major populations: SE dingoes, FI dingoes, NW dingoes and NGSD (Figure 4-7). Knowledge concerning mitochondrial and Y chromosome lineage of samples was plotted onto the phylogenetic tree (Figure 4-7, Appendix 2). To further explore the phylogenetic relationship between dingoes and NGSD a second rooted analysis was completed incorporating 12 wolf (Canis lupus) samples as outgroup taxa. SNPhylo pruned the

SNP dataset to 6,288 informative markers. The rooted ML analysis in raxmlGUI

(Silvestro & Michalak 2012) again observed that NGSD form their own monophyletic group compared to dingoes (Figure 4-8). Interestingly, rather than identifying a closer relationship between the NGSD and the SE dingo population as suggested in mtDNA analyses, the phylogenetic reconstructions suggest that the NGSD diverged before the dingo populations differentiated. Additionally, the whole genome SNP phylogeny indicates that the FI dingo population is closely related to the NW dingo populations with bootstrap support of 73 (60 in unrooted phylogeny), representing phylogenetic uncertainty. The split between the SE and NW dingo populations is strongly supported with a bootstrap value of 87 (100 in unrooted phylogeny).

145 ML analyses were also run with a topological constraint, forcing NGSD and SE dingoes to be monophyletic (HA) as in Chapter 2.Topological testing preformed in CONSEL

(Shimodaira and Hasegawa 2001a) rejected the alternative hypothesis (HA), supporting the hypothesis that NGSD form a monophyletic group distinct from dingoes (H0) (Table

4-3).

146 73 NGSD E 66 NGSD D

100 100 NGSD B NGSD C NGSD A Alpine 1 100 Alpine 5 Alpine 4 99 93 Alpine 2 Alpine 3 100 Fraser 7 Fraser 4 100 90 Fraser 5 Fraser 3 60 Simpson 1 Simpson 2 Kimberley 3 100 73 100 Northwestern 9 Kimberley 1 99 Kimberley 4 Kimberley 2 Simpson 5 91 Gibson 3 Gibson 2 100 Gibson 4

75 Gibson 5 Northwestern 2 Gibson 1

0.04 Figure 4-7 Maximum Likelihood tree based upon 4913 ancestry informative markers in

23 dingoes and 5 NGSD. The tree was constructed via the SNPhylo pipeline (Lee et al.

2014), with 6,000 non-parametric bootstrap replicates. Bootstrap values located above nodes, values below 70 not shown. Colours represent population clusters: red for SE dingoes, purple for FI dingoes, blue for NW dingoes and green for NGSD. Circles indicate mitochondrial lineage with; black for NW and orange for SE. Squares depict Y chromosome haplogroup with; green for H1, blue for H3 and red for H60 (Appendix 2).

147 Fraser 3 100 Fraser 7 75 Fraser 5 Fraser 4 Simpson 2 Simpson 5 82 73 94 Gibson 2 Gibson 3 100 Gibson 4 Gibson 1 Gibson 5 92 Northwestern 2 87 Kimberley 4 Kimberley 3 100 Kimberley 2 Kimberley 1 Northwestern 9 Simpson 1 Alpine 1 100 100 94 Alpine 5 79 Alpine 4 100 Alpine 2 Alpine 3 100 NGSD B NGSD C 100 NGSD E NGSD D NGSD A Wlf_LUb3 95 Wlf_LU1656 Wlf_19785 100 100 Wlf_22803 89 Wlf_22802 90 Wlf_LUb1 96 100 Wlf_LU1655 Wlf_22800 100 Wlf_22810 Wlf_22809 100 Wlf_LUb2 Wlf_LU1657 0.08 Figure 4-8 Maximum likelihood tree constructed based upon 6,288 informative SNPs in 23 dingoes, 5 NGSD and 12 wolves. The 12 wolf samples (Axelsson et al. 2013) were added as outgroup taxa. Tree constructed in raxmlGUI (Silvestro & Michalak

2012) using a GTR + G substitution model and 2000 bootstrap replicates. Bootstrap values located above nodes, values below 70 not shown. Colours represent population clusters: red for SE dingoes, purple for FI dingoes, blue for NW dingoes, green for

NGSD and gray for wolves.

148 Table 4-3 Topology testing between unconstrained and constrained maximum likelihood analyses. Calculated using CONSEL (Shimodaira and Hasegawa 2001a) Difference in log KH testΦ SH testΥ AU testΘ likelihood score p value p value p value Unconstrained vs 15.9 0.027 0.027 0.017 constrained treeΔ ΔConstrained topology forced the SE dingoes and NGSD to form a monophyletic group as observed in Chapter 2. ΦKishino-Hasegawa test (Kishino and Hasegawa 1989) ΥShimodaira-Hasegawa test (Shimodaira and Hasegawa 1999) ΘApproximately unbiased test (Shimodaira and Hasegawa 2001b)

149 4.4 Discussion

There are multiple lines of evidence that there are at least three genetically distinct dingo populations in Australia; SE dingoes, FI dingoes, NW dingoes, while NGSD form a separate distinct population. In clustering analyses, the SE and FI dingoes form discrete groups, whilst dingoes from the Kimberley, Simpson Desert and Gibson Desert cluster together in a NW population (Figure 4-2 and Figure 4-3). Principal components analysis and classical multidimensional scaling analyses also identified the presence of four population clusters (Figure 4-4-1, Figure 4-4-2 and Figure 4-5). However, whilst the NGSD, FI and NW clusters were tightly clustered, the SE samples were not. This suggests that the SE population is more heterogeneous than the other dingo and NGSD populations. This could be the result disturbed social structure due to higher levels of lethal control in southeastern Australia, leading to greater dispersal and mixing

(Wallach et al. 2009b). Or perhaps the population is ancestral or the heterogeneity could be the result of introgression from domestic dogs (Chapter 3).

FST values between the four population clusters suggest that gene flow between dingo populations is low, and that populations are highly divergent (Table 4-2). These data suggest that dingo populations may form discrete conservation units, and should ideally be managed separately. There was some evidence of admixture between the geographically subdivided populations, specifically from the FI population into the SE population and from the NW population into the SE population. There was also evidence of admixture from the SE and FI populations into the Simpson Desert (NW) dingoes. Dispersal is plausibly higher in southeastern Australia due to the disruption of

150 social structures through lethal control (Wallach et al. 2009b). Alternatively, these admixture patterns could be evidence of shared ancestral alleles or may indicate higher historical levels of introgression from domestic dogs (Chapter 3).

Phylogenetic and genetic distance analyses suggest that biogeographical patterns depicted in the whole genome SNP data (Figure 4-6, Figure 4-7 and Figure 4-8) are similar to the mitochondrial phylogeny and concatenated nuclear gene phylogeny

(Chapter 2), with two key differences. The presence of at least two dingo populations is consistent across all genetic markers (Figure 4-7, Figure 4-8, Chapter 2 and Chapter 3).

However, mitochondrial data suggested that SE and FI dingo populations are closely related whilst the whole genome SNP data, presented here, suggest that FI dingoes may be more closely related to the NW dingo populations (Figure 4-6, Figure 4-7, Figure 4-8 and Chapter 2).

Y chromosome haplotypes found in the FI population suggest that FI may have had paternal founders from the NW lineage (Chapter 3). In fact, FI dingoes predominantly shared Y chromosome haplotypes with dingoes from the NW populations (Chapter 3).

Bootstrap support for the FI/NW population grouping is low at 60-73% suggesting there is some uncertainty concerning the relationship of the FI population to the other dingo populations (Figure 4-7 and Figure 4-8). This uncertainty could be a reflection of the different evolutionary histories of the maternal, paternal and autosomal genetic markers.

One hypothesis to explain the results is that the FI population is the product of an initial foundation from the southeastern mitochondrial lineage, followed by paternal introgression from the northwestern lineages. It is possible that historical human

151 movements between mainland Australia and Fraser Island by Indigenous Australians may have facilitated historical paternal introgression. Clustering, phylogenetic, FST and

PCA analyses suggest that the FI dingo population currently forms a discrete population with little recent gene flow.

Mitochondrial (Chapter 2), Y chromosome (Chapter 3) and concatenated nuclear gene

(Chapter 2) analyses identified a close relationship between the NGSD and dingoes, as suggested by geographical proximity and previous genetic and morphological data

(Ardalan et al. 2012; Koler-Matznick et al. 2004; Oskarsson et al. 2011; Sacks et al.

2013; Savolainen et al. 2004). Mitochondrial and concatenated nuclear analyses in particular identified a closer relationship between NGSD and the SE and FI dingoes

(Chapter 2). Y chromosome data found that NGSD share the unique H60 SNP haplogroup with dingoes, and that single tandem repeat (STR) haplotypes were shared between NGSD and NW lineage dingoes (Chapter 3; Sacks et al. 2013).

Phylogenetic analyses based on whole genome SNP data, however, suggest that the

NGSD diverged before the two dingo lineages split (Figure 4-7 and Figure 4-8).

Topological testing and inclusion of outgroup taxa (wolves) further suggested that

NGSD form their own monophyletic lineage distinct from dingoes (Figure 4-8 and

Table 4-3). This suggests the relationship between NGSD and dingoes is not easily unravelled. Plausibly the discordancy between uniparental (Chapter 2 and Chapter 3) and whole genome phylogenies (Figure 4-7 and Figure 4-8) is the result of the different evolutionary histories depicted by different genetic markers. Discordancy between different genetic markers, particularly those with differences in inheritance,

152 recombination and effective population size are not uncommon (Galtier et al. 2009;

Humphries & Winker 2011; Rubinoff & Holland 2005; Toews & Brelsford 2012; Zink

& Barrowclough 2008). These results may also be a reflection of high inbreeding and genetic drift in the NGSD population, as the captive population was founded by a small number of individuals (Koler-Matznick et al. 2004). Possibly, these captive NGSD do not capture the full range of genetic variability present in wild NGSD. Further sampling of historical, wild or European captive NGSD across a range of genetic markers may help elucidate this relationship.

The maintenance of population structure in the absence of obvious physical barriers suggests that introduction history; behaviour and/or biology may be playing a role in shaping biogeographic patterns in the dingo. Mitochondrial and Y chromosome data suggests that these genetic populations may have diverged outside Australia (Chapter 2 and Chapter 3). Plausibly, the different genetic populations are the result of multiple introductions of dingoes into Australia, one spreading into the northwest and the other constrained to the southeast region of Australia (Chapter 2). The introductions may have occurred at different time points with the earlier population (southeastern) being swamped out in northern Australia. An alternative hypothesis is that the genetic populations are the result of adaptation to differences in habitat or ecological niches. In other carnivores, cryptic population subdivision has been observed along with differences in prey specialisation, dispersal patterns and habitat preference. (Musiani et al. 2007; Randall et al. 2010; Rueness et al. 2003a, 2003b; Sacks et al. 2004, 2008;

Stenseth et al. 2004; Vonholdt et al. 2010b; Weckworth et al. 2011). Mitochondrial and

153 Y chromosome data suggest that dispersal between the genetic populations is male- biased (Chapter 2 and Chapter 3).

Domestic dogs, specifically purebred domestic breeds, typically have elevated levels of inbreeding due to the effect of human artificial selection, population bottlenecks and line breeding (Leroy et al. 2009). FIS values for domestic dogs are reported to range between -0.2 and 0.19 (Fain et al. 2010; Huson et al. 2010; Leroy et al. 2009;

Ólafsdóttir & Kristjánsson 2008). Inbreeding in wild canid populations is reported to be lower than domestic dogs, with coyotes FIS = 0.04 (Fain et al. 2010) and wolves

FIS=0.06-0.08 (Fain et al. 2010; Kennedy et al. 1991; Stenglein et al. 2011). The SE and

Simpson Desert (NW) dingo populations have average FIS values similar to those of wild canid populations, whilst the Kimberley (NW) and Gibson Desert (NW) populations have slightly elevated average FIS levels (Table 4-1). Simpson Desert (NW) and SE dingo populations are under higher levels of human-mediated disruption, with ongoing management action plans possibly contributing to elevated dispersal/immigration events in these populations (Wallach et al. 2009b). It is possible that FIS levels are raised due to the ascertainment bias caused by phylogenetic distance between dingoes and the dog breeds used to develop the Illumina HD Canine Chip.

SNP array technologies have been successfully applied to other canids such as wolves and coyotes, with limited ascertainment bias observed (Boyko et al. 2010; vonHoldt et al. 2010a, 2011). Further sampling would be useful to investigate these patterns.

Inbreeding coefficients can be used to assess the genetic health of a population or species (Frankham & Ralls 1998; Keller & Waller 2002). High FIS levels may indicate

154 when a population is genetically unhealthy and at risk of inbreeding depression.

Inbreeding and resulting inbreeding depression can lead to the decline and/or extinction of the species or population (Ballou et al. 1998; Frankham & Ralls 1998; Keller &

Waller 2002; Laikre & Ryman 1991; O’Grady et al. 2006; Trinkel et al. 2011; Xu et al.

2007). An inbred Scandinavian wolf population on Isle Royale provides an example of how inbreeding depression can lead to increasing physiological anomalies including lower lifespan and elevated incidences of congenital defects (Räikkönen et al. 2013).

The inbreeding level of the Isle Royale wolves is F=0.27, approximately equal to offspring from sibling mating (Räikkönen et al. 2013). The extreme inbreeding levels observed in this study, FIS=0.647-0.732 (Table 4-1), suggest that the FI dingo population is not genetically healthy and the viability of the population is compromised.

However, the sample size is quite small and ascertainment bias could be affecting the

FIS levels, so further research is advised.

Current population estimates for dingoes on Fraser Island are approximately 100-200 individuals (Appleby & Jones 2011; DNPRSR 2014), down from approximately 300 in the 1990s (DASETT 1991). The FI dingo population is thought to serve as an important reservoir of pure dingoes isolated from mainland populations and relatively free from hybridisation (Stephens 2011; Woodall et al. 1996). The Fraser Island Dingo

Management Plan reports that the primary aim of the strategy is to “ensure conservation of a sustainable dingo population” (DEHP 2013). These data suggest a need for regular genetic monitoring on Fraser Island to track genetic diversity and possible inbreeding depression. Culls should not be made without further information regarding the genetic

155 health of the population which appears to be severely compromised and may require genetic rescue.

The FIS range observed in the NGSD, FIS=0.457-0.671, is also extremely high (Table 4-

1). NGSD in captive populations are likely experiencing high inbreeding due to a limited foundation population (Koler-Matznick et al. 2004). Introduction of new genetic lineages to the current NGSD population may be required to avoid inbreeding depression; however, this will likely prove difficult given NGSD have not been observed in the wild since the 1950s (Koler-Matznick et al. 2004).

A panel of 460 SNPs was identified as being variable and strongly ancestry informative.

These SNPs may prove useful for developing a specialised assay to investigate population structure in the dingo and NGSD across a larger sample size. In particular, incorporation of dingoes from northern Queensland, South Australia and the Northern

Territory will be important to more fully investigate the distribution of these dingo populations across Australia. Collection of historical and/or ancient DNA samples will also be useful for investigating introduction hypotheses and historical biogeography.

These data will be important for the implementation of conservation and management programs hoping to protect the genetic diversity and identity of the dingo.

Important future research directions may include increased sampling, both from new geographic regions and increased individuals from wild dingo and NGSD populations, to help elucidate the relationships between dingo populations and the NGSD. Dingoes from northern Queensland and historical samples may be important to investigate

156 historical biogeography in the dingo. Particularly important is the gathering of additional NGSD from either the wild or previously un-sampled captive lineages.

Another key research direction may be interrogating the presence of morphological and/or ecological differentiation between these genetic populations. Conservation groups have historically believed that there are 2-3 morphological varieties of dingo

(Corbett 1995; Walters 1995); here is evidence that there are at least three genetically subdivided dingo populations. A next step is to investigate the presence of morphological, biological or ecological differences between these genetic populations.

Comparison of dingoes to domestic dogs using whole genome SNP data may help elucidate markers for distinguishing dingoes from their hybrids. This is particularly an issue given that paternal introgression has been identified predominately in SE dingoes

(Chapter 3). Indeed two dingoes in this study carried H1 Y chromosome haplotypes, suggestive of European dog introgression (Figure 4-7, Appendix 2, Chapter 3). To investigate SNP markers useful in distinguishing dingoes, hybrids and domestic dogs further research incorporating dingoes, known hybrids and Australian domestic dogs will be necessary.

4.4.1 Conclusions

Whole genome SNP data indicate that there are at least three divergent dingo populations forming discrete conservation units. Gene flow is observed to low between the SE, FI and NW dingo populations. However, there is evidence from maternal and paternal markers that FI dingoes may be the product an initial foundation by SE dingoes

157 followed by successive paternal introgressions from the NW lineage. Historical human movements between mainland Australia and Fraser Island may have facilitated this paternal introgression pattern. NGSD form a discrete population closely related to dingoes. Concerningly, FI dingo populations and NGSD appear to be under threat of inbreeding. Management strategies may need to be adapted to preserve and improve the genetic health of these important canid populations. This study highlights the need to conserve dingoes from all geographical regions of Australia in order to preserve the full range of genetic diversity and identity of the dingo.

158 Chapter 5

General discussion

159 5.1 Introduction

Conservation groups and scientists have long debated the presence of phenotypic varieties of dingo, namely the alpine, desert and/or tropical. It is thought that the phenotypic varieties are distinguishable based upon morphological features such as coat type, size and colour (Walters 1995). The presence of these varieties, however, has not been rigorously scientifically investigated. This thesis presents compelling evidence from mitochondrial, Y chromosome, autosomal coding and whole genome single nucleotide polymorphism (SNP) markers that there are at least three populations of dingo: southeastern (SE), Fraser Island (FI) and northwestern (NW). The New Guinea

Singing Dog (NGSD) shares a close relationship with the dingo as evidenced by shared ancestry.

The presence of distinct dingo populations is an important finding, given current management and conservation strategies do not consider the presence of population subdivision. These three dingo populations should be treated as evolutionarily significant units distinct from each other. It is plausible that the different populations have divergent ecological profiles and are experiencing distinct stresses and pressures affecting their evolution.

5.2 Biogeography

Mitochondrial, Y chromosome and autosomal data presented in this thesis depict a pattern of geographical subdivision in the dingo, identifying three distinct dingo

160 populations: SE, FI and NW. Mitochondrial data suggests that maternal migration between populations is limited (Chapter 2 and Chapter 3). Y chromosome data, however, suggests that paternal introgression between populations is more common

(Chapter 3). This is not unexpected given ecological observations that male dingoes range further distances and are more likely to disperse than females (Thomson et al.

1992). Autosomal data only found limited evidence of introgression suggesting migration between populations is plausibly historical (Chapter 4). Plausibly the current distribution and ancestry of the two dingo lineages can be explained by a historical shift in their distribution and rare instances of gene flow due to ecological or behavioural barriers. Intriguingly Alpine 5, a sample collected in southeastern Australia seems to be evidence of historical maternal and paternal dispersal from the NW lineage into the SE lineage followed by successive backcrossing to the indigenous SE population. This hypothesis would explain the observation of a SE autosomal signature (Chapter 4) with

NW mitochondrial (Chapter 2) and Y chromosome (Chapter 3) ancestry within the same individual.

It is possible that the dingo populations have divergent ecological profiles. Patterns of genetic subdivision in other large carnivores such as: grey wolves, lynx, arctic foxes, orca and coyotes; have been observed to be associated with differences in ecology.

These ecological differences included prey specialisation (Carmichael et al. 2001;

Hoelzel et al. 1998; Munoz-Fuentes et al. 2007; Musiani et al. 2007, Pilot et al. 2006), dispersal/migration patterns (Dalén et al. 2005; McRae et al. 2005), neonatal habitat preference (Pilot et al. 2006; Rueness et al. 2003a, b; Sacks et al. 2004) and/or environmental climes (Rueness et al. 2003a, b). Evidence suggests that the different

161 dingo populations are likely the result of multiple introductions into Australia however it is plausible that the genetic lineages have different ecological profiles and/or specialisations. Potentially any differences in ecology and/or behaviour have contributed to the maintenance or establishment of genetic subdivision within the dingo.

Morphologically, dingoes and NGSD are similar but distinguishable (Crowther et al.

2014; Koler-Matznick et al. 2004). A close genetic relationship between dingoes and

NGSD is observed in the mitochondrial and Y chromosome data (Chapter 2 and

Chapter 3; Ardalan et al. 2012; Oskarsson et al. 2011; Sacks et al. 2013; Savolainen et al. 2004). Mitochondrial data suggest that dingoes from the SE and FI populations are more closely related to the NGSD than to the NW dingo population (Chapter 2 and

Chapter 3). Topology testing provides further evidence that dingoes are not monophyletic relative to the NGSD (Chapter 2). However Y chromosome haplotype sharing suggests that the NGSD and NW population are more closely related (Chapter

3). The discrepancy between maternal and paternal evolutionary histories suggests the relationship between dingoes and NGSD is complicated, and further research is required to unravel it.

Whole genome autosomal data suggests that NGSD form a discrete group outside of the dingo phylogeny (Chapter 4). This is contrary to mitochondrial (Chapter 2 and Chapter

3), nuclear coding (Chapter 2) and Y chromosome (Chapter 3) data that indicate that

NGSD lie within the dingo phylogeny. There are several hypotheses that could explain this discordancy. First NGSD samples are different between Chapters 2 and 3 and

Chapter 4, thus the divergent phylogenies could be a reflection of dissimilar

162 evolutionary signals from the different NGSD sample sets. An additional complication is that captive NGSD are highly inbred, thus this discrepancy could be a reflection of bias from inbreeding and random drift. Second, it is plausible that the evolutionary history of single autosomal genes (Chapter 2) and uniparental (Chapter 2 and Chapter 3) markers is divergent to the organismal or whole genome phylogeny (Chapter 4). Indeed, observations of discordancy between mitochondrial, autosomal and Y chromosome phylogenies are not infrequent (Galtier et al. 2009; Humphries & Winker 2011;

Rubinoff & Holland 2005; Toews & Brelsford 2012; Zink & Barrowclough 2008).

Third, this discrepancy could plausibly be the result of incomplete lineage sorting at mitochondrial and Y chromosome markers. Certainly these data suggest that the relationship between NGSD and dingoes is close but how close is yet to be resolved.

Further sampling of NGSD from the wild or additional captive lineages may further elucidate the relationship between dingoes and NGSD, as well as the evolutionary history of dogs in Oceania. Integration of additional South East Asian dogs, particularly pre-Neolithic samples, may help further clarify the relationship between dingoes and

NGSD.

5.3 Introduction route, origin and arrival

There are a range of hypotheses concerning the origin of dingoes (Figure 5-1). The data collected in this thesis corroborate previous studies suggesting that dingoes have an

Asian origin. Dingo mitochondrial and Y chromosome haplotypes are most closely related to those observed in South East Asian dogs (Chapter 2 and Chapter 3; Ardalan et al. 2012; Brown et al. 2011; Oskarsson et al. 2011; Sacks et al. 2013; Savolainen et al.

163 2004). There is, however, limited haplotype sharing between Oceania and South East

Asia, suggesting shared ancestry but long-term isolation between these canine populations. Nuclear data also suggests that dingoes share an affinity with Asian wolves and dogs (vonHoldt et al. 2010). Mitochondrial haplotypes observed in dingoes are not similar to those observed in Indian dogs by Pang et al (2009), indicating an Indian origin as suggested by Pugach et al. (2013) is unlikely. The close relationship with

NGSD suggests that dingoes and NGSD share ancestry. The most likely introduction route for dingoes was through Papua New Guinea, which was joined to Australia through a land bridge 6,000-8,000 years ago (Figure 5-1). An expectation of this hypothesis is that wild or historical NGSD would exhibit higher autosomal heterogeneity given the possible convergence of two ancestral lineages in Papua New

Guinea.

Savolainen et al. (2004) suggest dingoes are the result of a single small homogeneous introduction, possibly even a single pregnant female. However, the presence of at least two distinct mitochondrial and Y chromosome lineages with no intermediate haplotypes found in Oceania suggests that dingoes are more likely the result of multiple immigrations into Australia (Chapter 2 and Chapter 3). Mitochondrial relationships between dingoes and NGSD suggest the dingo populations diverged outside of

Australia. Y chromosome haplotypes observed in the dingo populations are also divergent and suggestive of a split outside Australia. Specifically, one Y haplogroup,

H60, shares a relationship with H5, a haplogroup only observed in Taiwan, whilst the other, H3, is shared with South East Asian dogs (Ardalan et al. 2012; Sacks et al. 2013).

164 Mitochondrial divergence estimates suggest that plausibly the populations immigrated at different times (Chapter 2).

Mitochondrial-dating efforts should be considered cautiously since discordancy has been observed between mitochondrial and whole genome phylogenetic reconstructions

(Chapter 2 and Chapter 4). However, several lines of evidence corroborate the antiquity of dingoes. The dingo is an ancient dog (vonHoldt et al. 2010). The oldest fossilised dingo remains are 3,500 years old; suggesting the minimum arrival time of dingo was approximately 4,000 years ago (Gollan 1984; Macintosh 1964, 1975). Molecular dating estimates from mitochondrial data suggest the mitochondrial lineages diverged approximately 10,000 years ago (Chapter 2). Prior molecular dating estimates based on the mitochondrial control region also suggest an older mitochondrial divergence time is possible (Oskarsson et al. 2011). Y chromosome studies found evidence that dingoes are from an older dog radiation than dogs in South East Asia (Sacks et al. 2013).

Amylase copy number investigations observed that dingoes only have the ancestral two copies, suggesting that they diverged from domestic dogs before agriculture arose

(Axelsson et al. 2013; Freedman et al. 2014; Cairns, unpublished data).

165 30 20 10 0 −10 −20 −30 −40

90 100 110 120 130 140 150

Figure 5-1 Map depicting hypotheses of the origin and introduction route of the dingo.

Black represents the hypothesis presented by Gollan (1984) and Pugach et al. (2013) postulating origination from India. Green represents the hypotheses presented by

Corbett (1995), Savolainen (2004) and Oskarsson et al. (2011) with dingoes originating in mainland Asia. Blue represents Sacks et al. (2013) postulating origination in Taiwan.

Red represents the hypotheses presented in this thesis, i.e. multiple South East Asian origins and introduction through Papua New Guinea.

166 Several lines of evidence suggest that dingoes were not brought to Australia as the result of the Neolithic human expansion; lack of agriculture, no cultural items such as pigs or chickens and no observation of gene flow from Neolithic humans are suggestive of this (Bocquet-Appel 2011; Bramanti et al. 2009; Haak et al. 2010; Karafet et al.

2005; Larson et al. 2007; McEvoy et al. 2010; Mulvaney & Kamminga 1999;

Oskarsson et al. 2011; Sacks et al. 2013; Chapter 2). Together, these data support the hypothesis that dingoes plausibly arrived in Australia earlier than thought, and certainly at a minimum of 5,000 years ago. This is at odds with the current dogma that dingoes arrived in Australia as part of the Neolithic human expansion (Corbett 1995; Savolainen et al. 2004).

5.4 The mixed identity of Fraser Island dingoes

MtDNA and Y chromosome data observed the presence of two subdivided populations of dingo, however, whole genome SNP data provide evidence that FI is a third distinct population (Chapter 2, Chapter 3 and Chapter 4). Divergent maternal and paternal histories indicate that the FI dingo population may be the result of mixing between lineages (Chapter 2 and Chapter 3). Autosomal SNP data provide evidence that gene flow between mainland dingo populations and FI dingoes is limited, demonstrating that

FI dingoes are likely a discrete population (Chapter 4). Mitochondrial data indicates that the FI dingo population is fairly homogeneous with all individuals carrying SE lineage haplotypes (Chapter 2 and Chapter 3). Y chromosome data, however, suggested at least four distinct paternal lineages, two H60, an H3 and an H1 haplotype, are present on

Fraser Island (Chapter 3). This seems logically consistent with an initial foundation by

167 SE population dingoes, followed by successive paternal introgressions plausibly from the NW dingo population. Indigenous Australians, probably accompanied by dingoes, have lived on Fraser Island for at least 3,000 years (McNiven 1998). It is possible that human movements between Fraser Island and the Australian mainland are responsible for the observed pattern of paternal gene flow into FI dingoes.

Fraser Island is generally accepted to be an important reservoir for pure dingoes, with hybridisation levels reported to be low (Stephens 2011; Woodall et al. 1996). These data, however, suggest that dingoes on Fraser Island are not genetically representative of dingoes from the mainland (Chapter 3 and Chapter 4). Conservation programs should not be reliant on the FI dingoes as a source of genetic diversity representative of the mainland dingoes; they are distinct evolutionarily significant units.

Autosomal single nucleotide polymorphism (SNP) data collected in Chapter 4 suggests that inbreeding has compromised the FI dingo population. It is unlikely that the FI dingo population has been historically maintained at such high inbreeding levels; instead it is likely the result of recent demographic shifts and low gene flow from the

Australian mainland. FI dingo populations are believed to be declining. No proper census of dingo population numbers on Fraser Island have been completed, however, population estimates in the 1990s suggested there were approximately 300 individuals

(DASETT 1991) and in 2011 there were approximately 150 individuals (Appleby &

Jones 2011). A population becomes unsustainable when the effective population size drops below 100-200 (Ballou et al. 1998; Keller & Waller 2002). Inbreeding levels observed in the FI population suggest it may be at risk of inbreeding depression. In a

168 wolf population on Isle Royale with similarly high inbreeding levels, increasing levels of congenital deformities and shorter lifespans have been observed (Räikkönen et al.

2013). There is no published data concerning the incidences of congenital deformities, reproduction success or lifespan for dingoes on Fraser Island, and this represents a severe knowledge gap. Management of the Fraser Island population has been extremely controversial with culls being performed to limit negative human-dingo interactions

(DEHP 2013). However, the primary aim of the Queensland state government management strategies is to “ensure conservation of a sustainable dingo population”

(DEHP 2013; DNPRSR 2014). In light of data from this thesis, more research into the genetic and physiological health of the FI dingo population is vital.

5.5 Hybridisation – evidence of historical paternal introgression

Hybridisation between dingoes and European domestic dogs is a significant problem for the conservation and management of dingoes. Morphological and genetic surveys have observed that in the northwest of Australia a large proportion of the dingoes are genetically intact whilst in southeastern Australia dingoes free from domestic dog introgression are less common (Claridge et al. 2009; Corbett 2001; Jones 2009;

Newsome & Corbett 1982, 1985; Newsome et al. 1980; Robley et al. 2010; Stephens

2011; Wilton 2001; Woodall et al. 1996). In the 1990s a genetic test was developed by

Wilton and colleagues to estimate the genetic integrity or purity of a dingo individual

(Wilton 2001; Wilton et al. 1999). This was an improvement on earlier methods that required measurements of skulls from deceased animals, or classifications based on unreliable indicators like coat colour (Corbett 1995; Elledge et al. 2008; Newsome et al.

169 1980). Mitochondrial control region studies suggested that the genetic testing method was fairly robust, with only two out of 326 dingoes carrying non-dingo mitochondrial haplotypes (Oskarsson et al. 2011; Sacks et al. 2013; Savolainen et al. 2004). In

Chapters 2 and 3, little evidence of domestic dog introgression from maternal markers was observed, with only a single dog-like mitochondrial type being identified. However in Chapter 3, evidence of male domestic dog introgression into dingoes was observed.

Specifically, the presence of Y chromosome SNP and STR haplotypes in dingoes that are shared with modern European domestic dog breeds, such as Labradors, which is strongly suggestive of hybridization rather than independent mutation (Chapter 3, Sacks et al. 2013). This hints that domestic dog introgression into dingoes is male biased with most hybridisation occurring between female dingoes and male domestic dogs.

Evidence of introgression was predominately localised to the SE dingo population.

Additionally, it suggests that the SE dingo population is particularly at risk of extinction through introgression. This is corroborated by a large-scale survey of introgression levels in approximately 4,500 wild dingoes by Stephens (2011), who found that hybridisation is relatively low in outback Australia but substantial in southeastern

Australia.

In Chapter 4, two of the male dingoes carried H1 Y chromosome haplotypes suggesting the presence of paternal domestic dog introgression (Chapter 4, Appendix 2).

Mitochondrial and autosomal markers failed to detect this introgression. This indicates that likely extensive backcrossing with wild dingoes occurred following any historical paternal hybridisation making it difficult to observe introgression at autosomal loci

(Chapter 2, Chapter 3, Chapter 4). Utilisation of the whole genome SNP dataset to

170 compare dingoes and domestic dogs may prove useful to investigate markers capable of distinguishing between dingoes, hybrids and domestic dogs (Chapter 1). However, given that the Illumina canine HD SNP chip was designed to incorporate only SNPs observed to be variable in domestic dogs a better method may be to collect whole genome sequence data for comparison to domestic dogs to elucidate the presence of fixed diagnostic SNPs between dingoes and dogs (Chapter 1). This data also suggests the need for integration of additional markers to the dingo genetic test; specifically, the

Y chromosome markers utilized in Chapter 3 would be useful to detect paternal introgression.

Lethal control programs are common in southeastern Australia due to livestock agricultural activities (Corbett 1995; Downward & Bromell 1990; Fleming et al. 2001,

2006). Current lethal management strategies are plausibly exacerbating hybridisation in southeastern Australia due to the breakdown of dingo social structures possibly increasing breeding events between dingoes and domestic dogs (Wallach et al. 2009b).

Management strategies need to be adapted to limit hybridisation and preserve existing genetically intact SE dingo populations.

5.6 Future of the dingo –conservation and sustainable management

In the past it was considered that dingoes formed a single homogeneous population

(Savolainen et al. 2004). However, the combined data collected in this thesis suggest that dingoes are not a single homogeneous population, but rather at least three distinct populations or evolutionarily significant units. This new knowledge concerning cryptic

171 population structure in the dingo presents a number of conservation and management considerations. First, captive breeding programs should endeavour to limit mixing between the distinct genetic populations. Genetic testing methodologies could be adapted to help inform these decisions by incorporating variable mitochondrial, Y chromosome and autosomal SNP markers, as identified in this thesis, that distinguish between genetic populations (Chapter 2, Chapter 3 and Chapter 4). Second, national management and conservation strategies need to be adapted to preserve the distinct dingo populations. Current management strategies consider that dingoes south of the

“dingo fence” should be eradicated outside of national parks (Fleming et al. 2001). This significantly endangers the SE dingo population, an evolutionarily significant population. Strategies should endeavour to collect and use genetic data concerning the identity of local dingoes and manage the SE, FI and NW populations as distinct entities.

Limited gene flow between the populations, as observed by high FST values (Chapter 2,

Chapter 3 and Chapter 4), suggests human-mediated translocation between the three distinct dingo populations is not recommended. Large-scale genetic surveys may be needed to inform management decisions, particularly in regions that were not sampled in these studies. In Chapter 4, a set of 460 autosomal SNP markers was identified that may prove particularly useful for future genetic surveys aiming to investigate biogeographical variation in the dingo.

Hybridisation plays a significant role in current management strategies. The importance of the SE dingo population suggests a shift in thinking is required, despite higher levels of hybridisation. This is not a threat unique to dingoes; other wild species such as the

American Bison (Bison bison) and wildcats (Felis silvestris) with domesticated relatives

172 also face introgression pressures (Daniels & Corbett 2003; Freese et al. 2007; Halbert

2003; Halbert & Derr 2007; Hertwig et al. 2009; Ward et al. 1999). A “hybrid reduction” strategy is one management option, where hybrids with detectable introgression (or an introgression level above a certain level) are removed from the breeding population (Halbert 2003). This strategy will not re-create a genetically “pure” population but will serve to reduce the level of hybridisation. A controversial study by

Garrick et al. (2012) suggests that targeted backcrossing with individuals of high genetic integrity individuals may allow the reconstruction of a highly threatened species.

Legislation concerning the dingoes’ status differs between and within Australian states and territories (Davis 2001; Downward & Bromell 1990; Fleming et al. 2001, 2006;

Trigger et al. 2008). Legislation needs to be revised to reflect the need for preservation of the identified genetically distinct populations. Legislation should also reflect the vital role dingoes play in Australia’s ecological framework as trophic regulators (Dickman et al. 2009; Fillios et al. 2010; Glen & Dickman 2005; Glen et al. 2007; Johnson et al.

2007; Letnic & Crowther 2012; Letnic & Koch 2010; Pople et al. 2000; Wallach et al.

2010), indirect protection to small marsupials (Johnson et al. 2007; Letnic et al. 2009a,

2009b; Wallach et al. 2009a) and in controlling invasive mesopredators (Cupples et al.

2011; Johnson & VanDerWal 2009; Kennedy et al. 2012; Letnic et al. 2009a, 2010;

Letnic & Dworjanyn 2011; Letnic & Koch 2010; Moseby et al. 2012; Wang & Fisher

2012). Evidence that lethal control measures are particularly damaging to dingoes by degrading social structures suggests alternative measures need to be investigated to help balance the conservation of the dingo and agricultural interests (Wallach et al. 2009b).

173 There is also increasing evidence that lethal control may be counter-productive for cattle producers (Allen 2014) and in controlling dingo population sizes (Thomson 1992;

Wallach et al. 2009b, 2010). Non-lethal control measures such as adequate fencing, guardian animals and perimeter scents may prove to be viable options (Fleming et al.

2001, 2006; van Bommel & Johnson 2012). Such measures have been successful in the

United States where coyotes present a similar risk to livestock (Green et al. 1984;

Shivik 2004, 2006).

5.7 Future directions

There are still several key knowledge gaps concerning dingoes, including the arrival time and route of the dingo, the relationship between dingoes and the NGSD. Also the genetic health of the FI dingo population, the genetic identity of dingoes in un-sampled regions of Australia and the presence of morphological or ecological variation between the dingo populations need to be investigated. Investigations into the time and route of the dingoes’ arrival into Australia might aim to incorporate a range of genetic markers as well as historical sampling. Use of historical samples from both Australia and Papua

New Guinea may prove particularly important given the close relationship between dingoes and NGSD. To elucidate the relationship between dingoes and NGSD, a wider range of NGSD samples is needed, either from the wild, historical collections, or additional captive lineages. The genetic health of the Fraser Island dingo population is a particularly pressing concern; a widespread genetic survey along with ecological data is needed to assess inbreeding levels and potential inbreeding depression. A survey of dingoes from previously un-sampled and conservation significant populations is needed

174 to continue informing management and conservation programs. Also of interest may be the presence of morphological, ecological or behavioural variation between the genetically distinct dingo populations that could be contributing to or maintaining reproductive isolation.

5.8 Conclusions

Dingoes are not a single homogeneous population, instead there are at least three evolutionarily significant populations: the southeastern, Fraser Island and northwestern populations. The different populations are likely experiencing different evolutionary pressures, with the southeastern population at high risk of extinction through hybridisation and the Fraser Island population possibly compromised by high inbreeding levels. Current conservation and management strategies need to be adapted to reflect the presence of population structure in the dingo. National and state legislation also need to better reflect the presence of these three evolutionarily significant dingo populations, as well as, the dingoes’ important ecological role in Australia. Further genetic studies may aim to link morphological and/or ecological variation with observed genetic subdivision and further explore the evolution of the dingo and New Guinea

Singing Dog.

175 References

176 Chapter 1 – 5 references

Adler D, Murdhoch D (2014) rgl: 3D visualization device system (OpenGL). Aggarwal RK, Kivisild T, Ramadevi J, Singh L (2007) Mitochondrial DNA coding region sequences support the phylogenetic distinction of two Indian wolf species. Journal of Zoological Systematics and Evolutionary Research 45, 163- 172. Alexander DH, Lange K (2011) Enhancements to the ADMIXTURE algorithm for individual ancestry estimation. BMC bioinformatics 12, 246. Alexander DH, Novembre J, Lange K (2009) Fast model-based estimation of ancestry in unrelated individuals. Genome Research 19, 1655-1664. Allen BL (2010) Did dingo control cause the elimination of kowaris through mesopredator release effects? A response to Wallach and O'Neill (2009). Animal Biodiversity and Conservation 33, 205-208. Allen BL, Engeman RM, Allen LR (2011a) Wild dogma II: The role and implications of wild dogma for wild dog management in Australia. Current Zoology 57, 737- 740. Allen BL, Engeman RM, Allen LR (2011b) Wild dogma: An examination of recent "evidence" for dingo regulation of invasive mesopredator release in Australia. Current Zoology 57, 568-583. Allen BL, Fleming PJS (2012) Reintroducing the dingo: the risk of dingo predation to threatened vertebrates of western New South Wales. Wildlife Research 39, 35- 50. Allen LR (2014) Wild dog control impacts on calf wastage in extensive beef cattle enterprises. Animal Production Science 54, 214-220. Anderson TM, vonHoldt BM, Candille SI, et al. (2009) Molecular and evolutionary history of melanism in North American Gray Wolves. Science 323, 1339-1343. Anonnymous (2010) Marchers to rally for threatened Fraser dingoes. In: Sydney Morning Herald. Fairfax Publishing, Sydney. Appleby R, Jones D (2011) Analysis of a preliminary dingo capture- mark-recapture experiment on Fraser Island conducted by the Queensland Parks and Wildlife Service (ed. Service QPaW), Queensland, Australia. Ardalan A, Kluetsch CFC, Zhang A-b, et al. (2011) Comprehensive study of mtDNA among Southwest Asian dogs contradicts independent domestication of wolf, but implies dog–wolf hybridisation. Ecology and Evolution 1, 373-385. Ardalan A, Oskarsson M, Natanaelsson C, et al. (2012) Narrow genetic basis for the Australian dingo confirmed through analysis of paternal ancestry. Genetica 140, 65-73. Axelsson E, Ratnakumar A, Arendt M-L, et al. (2013) The genomic signature of dog domestication reveals adaptation to a starch-rich diet. Nature 495, 360-364. Ballard JWO, Whitlock MC (2004) The incomplete natural history of mitochondria. Molecular Ecology 13, 729-744. Ballou JD, Breitenmoser-Würsten C, Rohner C, Breitenmoser U (1998) General problems of small (carnivore) populations: minimum viable population size and inbreeding. Strasbourg: Council of Europe Publishing Environmental encounters 38, 27-40. Bandelt H-J, Forster P, Röhl A (1999) Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16, 37-48.

177 Behrensmeyer AK (1978) Taphonomic and ecologic information from bone weathering. Paleobiology 4, 150-162. Berryere TG, Kerns JA, Barsh GS, Schmutz SM (2005) Association of an Agouti allele with fawn or sable coat color in domestic dogs. Mammalian Genome 16, 262- 272. Bocquet-Appel J-P (2011) When the world’s population took off: the springboard of the Neolithic demographic transition. Science 333, 560-561. Boyko AR, Quignon P, Li L, et al. (2010) A simple genetic architecture underlies morphological variation in dogs. PLoS Biology 8, e1000451. Bramanti B, Thomas MG, Haak W, et al. (2009) Genetic discontinuity between local Hunter-Gatherers and central Europe’s first farmers. Science 326, 137-140. Breckwoldt R (1988) A very elegant animal, The Dingo. Angus and Robertson Publishers, Sydney. Brito P, Edwards S (2009) Multilocus phylogeography and phylogenetics using sequence-based markers. Genetica 135, 439-455. Brook LA, Johnson CN, Ritchie EG (2012) Effects of predator control on behaviour of an apex predator and indirect consequences for mesopredator suppression. Journal of Applied Ecology 49, 1278-1286. Brown M (2010) Yes, a dingo could have killed the baby. In: Sydney Morning Herald. Fairfax Publishing, Sydney. Brown RP, Yang Z (2011) Rate variation and estimation of divergence times using strict and relaxed clocks. BMC Evolutionary Biology 11, 271. Brown SK, Pedersen NC, Jafarishorijeh S, et al. (2011) Phylogenetic distinctiveness of Middle Eastern and Southeast Asian village dog Y chromosomes illuminates dog origins. PLoS ONE 6, e28496. Brownrigg R, Minka TP, Becker RA, Wilks AR (2014) maps: Draw Geographical Maps. Bueler LE (1974) Wild dogs of the World. Constable and Company Limited, London. Bultman SJ, Klebig ML, Michaud EJ, et al. (1994) Molecular analysis of reverse mutations from non-agouti (a) to black-and-tan (at) and white-bellied agouti (aw) reveals alternative forms of agouti transcripts. Genes & Development 8, 481-490. Butchart SHM, Walpole M, Collen B, et al. (2010) Global biodiversity: Indicators of recent declines. Science 328, 1164-1168. Cairns HC, Harney BY (2004) Dark Sparklers. H. C. Cairns, Merimbula. Cadieu E, Neff MW, Quignon P, et al. (2009) Coat variation in the domestic dog is governed by variants in three genes. Science 326, 150-153. Candille S (2009) Stanford Brindle Study Report. Great Dane Club of Western New York inc. http://www.greatdaneclubofwny.com/educational mat.htm Candille SI, Kaelin CB, Cattanach BM, et al. (2007) A -Defensin mutation causes black coat color in domestic dogs. Science 318, 1418-1423. Candille SI, Van Raamsdonk CD, Chen C, et al. (2004) Dorsoventral patterning of the mouse coat by Tbx15. PLoS Biology 2, 0030-0042. Catling PC, Corbett LK, Newsome AE (1992) Reproduction in captive and wild dingoes (Canis familiaris dingo) in temperate and arid environments of australia. Wildlife Research 19, 195-209.

178 Carmichael LE, Nagy JA, Larter NC, Strobeck C (2001) Prey specialization may influence patterns of gene flow in wolves of the Canadian Northwest. Molecular Ecology 10, 2787-2798. Carthey AJR, Banks PB (2012) When does an alien become a Native species? A vulnerable Native mammal recognizes and responds to its long-term alien predator. PLoS ONE 7, e31804. Caughley G, Grigg GC, Caughley J, Hill GJE (1980) Does dingo predation control the densities of kangaroos and emus? Wildlife Research 7, 1-12. Chase K, Jones P, Martin A, Ostrander EA, Lark KG (2009) Genetic mapping of fixed phenotypes: Disease frequency as a breed characteristic. Journal of Heredity 100, S37-41. Claridge AW, Mills DJ, Hunt R, Jenkins DJ, Bean J (2009) Satellite tracking of wild dogs in south-eastern mainland Australian forests: implications for management of a problematic top-order carnivore. Forest Ecology and Management 258, 814-822. Clement M, Posada D, Crandall K (2000) TCS: a computer program to estimate gene genealogies. Molecular Ecology 9, 1657. Corbett LK (1988) Social dynamics of a captive dingo pack: Population regulation by dominant female infanticide. Ethology 78, 177-198. Corbett LK (1995) The dingo in Australia and Asia. University of NSW Press, Sydney. Corbett LK (2001) Conservation status of the dingo. In: A symposium on the dingo (eds. Dickman CR, Lunney D), pp. 10-19. Royal Zoological Society of New South Wales, Sydney. Corbett LK, Newsome AE (1975) Dingo Society and its maintenance: a preliminary analysis. In: The Wild Canids - their systematics, behavioural ecology and evolution (ed. Fox MW), pp. 369-379. Van Nostrand Reinhold Company, New York. Crowther MS, Fillios M, Colman N, Letnic M (2014) An updated description of the Australian dingo (Canis dingo Meyer, 1793). Journal of Zoology 293, 192-203. Cupples JB, Crowther MS, Story G, Letnic M (2011) Dietary overlap and prey selectivity among sympatric carnivores: could dingoes suppress foxes through competition for prey? Journal of Mammalogy 92, 590-600. Cyranoski D (2005) Japan jumps towards personalised medicine. Nature 437, 796. Dalén L, Fuglei EVA, Hersteinsson P, et al. (2005) Population history and genetic structure of a circumpolar species: the arctic fox. Biological Journal of the Linnean Society 84, 79-89. Daniels MJ, Corbett L (2003) Redefining introgressed protected mammals: when is a wildcat a wild cat and a dingo a wild dog? Wildlife Research 30, 213-218. Darriba D, Taboada GL, Doallo R, Posada D (2012) jModelTest 2: more models, new heuristics and parallel computing. Nature Methods 9, 772-772. DASETT (1991) Nomination of Fraser Island and the Great Sandy Region by the government of Australia for inclusion in the World Heritage List. Department of the Arts, Sport, the Environment, Tourism and the Territories, Canberra, Australia. David B, Barker B, Petchey F, et al. (2013) A 28,000 year old excavated painted rock from Nawarla Gabarnmang, northern Australia. Journal of Archaeological Science 40, 2493-2501.

179 Davis E (2001) Legislative issues relating to control of dingoes and other wild dogs in New South Wales. 1. Approaches to future management. In: A symposium on the Dingo (eds. Dickman CR, Lunney D). Royal Zoological Society of New South Wales, Sydney. Degnan JH, Rosenberg NA (2009) Gene tree discordance, phylogenetic inference and the multispecies coalescent. Trends in Ecology & Evolution 24, 332-340. DEHP (2013) Fraser Island Dingo Conservation and Risk Management Strategy (ed. Ecosystem Services DoEaHP). Queensland Government, Queensland, Australia. Diamond J (2002) Evolution, consequences and future of plant and animal domestication. Nature 418, 700-707. Dickman CR (1996) Impact of exotic generalist predators on the native fauna of Australia. Wildlife Biology 2, 185-195. Dickman CR, Glen AS, Letnic M (2009) Reintroducing the dingo: Can Australia's conservation wastelands be restored? In: Reintroduction of Top-Order Predators, pp. 238-269. Wiley-Blackwell. Ding ZL, Oskarsson M, Ardalan A, et al. (2011) Origins of domestic dog in Southern East Asia is supported by analysis of Y-chromosome DNA. Heredity 108, 507- 514. DNPRSR (2014) About Fraser Island dingoes: Myths and realities (FAQs) (ed. Department of National Parks R, Sport and Racing). Queensland Government, Queensland, Australia. Downward RJ, Bromell JE (1990) The development of a policy for the management of dingo populations in South Australia. In: Proceedings of the Fourteenth Vertebrate Pest Conference 1990. Dreger DL, Schmutz SM (2010) A new mutation in MC1R explains a coat color phenotype in 2 "old" breeds: Saluki and Afghan Hound. Journal of Heredity 101, 644-649. Drummond AJ, Suchard MA, Xie D, Rambaut A (2012) Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution 29, 1969-1973. Druzhkova AS, Thalmann O, Trifonov VA, et al. (2013) Ancient DNA analysis affirms the canid from Altai as a primitive dog. PLoS ONE 8, e57754. Drögemüller C, Giese A, Martins-Wess F, et al. (2006) The mutation causing the black- and-tan pigmentation phenotype of Mangalitza pigs maps to the porcine ASIP locus but does not affect its coding sequence. Mammalian Genome 17, 58-66. Edwards SV (2009) Is a new and general theory of molecular systematics emerging? Evolution 63, 1-19. Edwards SV & Beerli P (2000) Perspective: Gene divergence, population divergence, and the variance in coalescence time in phylogeographic studies. Evolution 54, 1839-1854. Eldridge SR, Shakeshaft BJ, Nano TJT (2002) The impact of wild dog control on cattle, native and introduced herbivores and introduced predators in central Australia, Final report to the Bureau of Rural Sciences. (ed. Territory PaWCotN), Alice Springs Elledge AE, Allen LR, Carlsson B-L, Leung LK-P (2008) An evaluation of genetic analyses, skull morphology and visual appearance for assessing dingo purity: implications for dingo conservation. Wildlife Research 35, 812-820.

180 Elledge AE, Leung LK-P, Allen LR, Firestone K, Wilton A (2006) Assessing the taxonomic status of dingoes Canis familiaris dingo for conservation. Mammal Review 36, 142-156. Endicott P, Ho SYW (2008) A Bayesian evaluation of Human mitochondrial substitution rates. The American Journal of Human Genetics 82, 895-902. Everts RE, Rothuizen J, van Oost BA (2000) Identification of a premature stop codon in the melanocyte-stimulating hormone receptor gene (MC1R) in Labrador and Golden retrievers with yellow coat colour. Animal Genetics 31, 194-199. Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10, 564-567. Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to Human mitochondrial DNA restriction data. Genetics 131, 479-491. Fain SR, Straughan DJ, Taylor BF (2010) Genetic outcomes of wolf recovery in the western Great Lakes states. Conservation Genetics 11, 1747-1765. Felsenstein J (2005) PHYLIP (Phylogeny Inference Package) version 3.6, Department of Genome Sciences, University of Washington, Seattle. Fillios M, Crowther MS, Letnic M (2012) The impact of the dingo on the thylacine in Holocene Australia. World Archaeology 44, 118-134. Fillios M, Gordon C, Koch F, Letnic M (2010) The effect of a top predator on kangaroo abundance in arid Australia and its implications for archaeological faunal assemblages. Journal of Archaeological Science 37, 986-993. Fleming P, Corbett LK, Harden R, Thomson P (2001) Managing the impacts of dingoes and other wild dogs. Bureau of Rural Sciences, Canberra. Fleming PJS, Allen BL, Ballard G-A (2012a) Cautionary considerations for positive dingo management: a response to the Johnson and Ritchie critique of Fleming et al. (2012). Australian Mammalogy 35, 15-22. Fleming PJS, Allen BL, Ballard G-A (2012b) Seven considerations about dingoes as biodiversity engineers: the socioecological niches of dogs in Australia. Australian Mammalogy 34, 119-131. Fleming PJS, Allen LR, Lapidge SJ, et al. (2006) A strategic approach to mitigating the impacts of wild canids: proposed activities of the Invasive Animals Cooperative Research Centre. Australian Journal of Experimental Agriculture 46, 753-762. Forster P, Röhl A, Lünnemann P, Brinkmann C et al. (2000) A short tandem repeat based phylogeny for the human Y chromosome. The American Journal of Human Genetics 67, 182–196. Frankham R, Ralls K (1998) Conservation biology: Inbreeding leads to extinction. Nature 392, 441-442. Freedman AH, Gronau I, Schweizer RM, et al. (2014) Genome sequencing highlights the dynamic early history of dogs. PLoS Genetics 10, e1004016. Freese CH, Aune KE, Boyd DP, et al. (2007) Second chance for the plains bison. Biological Conservation 136, 175-184. Fu YX (1997) Statistical Tests of neutrality of mutations against population growth, hitchhiking and background Selection. Genetics 147, 915-925. Fu YX, Li WH (1993) Statistical tests of neutrality of mutations. Genetics 133, 693- 709.

181 Gadagkar SR, Rosenberg MS, Kumar S (2005) Inferring species phylogenies from multiple genes: Concatenated sequence tree versus consensus gene tree. Journal of Experimental Zoology Part B: Molecular and Developmental Evolution 304, 64-74. Galtier N, Nabholz B, GlÉMin S, Hurst GDD (2009) Mitochondrial DNA as a marker of molecular diversity: a reappraisal. Molecular Ecology 18, 4541-4550. Garrick RC, Benavides E, Russello MA, et al. (2012) Genetic rediscovery of an ‘extinct’ Galápagos giant tortoise species. Current Biology 22, R10-R11. Gibbons A (2001) The peopling of the Pacific. Science 291, 1735 - 1737. Glen AS (2010) Hybridisation between dingoes and domestic dogs: a comment on Jones (2009). Australian Mammalogy 32, 76-77. Glen AS (2012) Enough dogma: Seeking the middle ground on the role of dingoes. Current Zoology 58, 856-858. Glen AS, Dickman CR (2005) Complex interactions among mammalian carnivores in Australia and their implications for wildlife management. Biological Reviews of the Cambridge Philosophical Society 80, 387-401. Glen AS, Dickman CR, Soulè ME, Mackey BG (2007) Evaluating the role of the dingo as a trophic regulator in Australian ecosystems. Austral Ecology 32, 492-501. Gollan K (1984) The Australian dingo: in the shadow of man. In: Vertebrate Zoogeography & Evolution in Australasia (eds. Archer M, Clayton G). Hesperian Press, Carlisle, WA. Gould J (1863) The Mammals of Australia volume III. MacMillan Company, Melbourne. Gray S (2010) Tourists 'threatening dingo exinction'. In: Sydney Morning Herald. Fairfax Publishing, Sydney. Green JS, Woodruff RA, Tueller TT (1984) Livestock-guarding dogs for predator control: costs, benefits, and practicality. Wildlife Society Bulletin, 44-50. Haak W, Balanovsky O, Sanchez JJ, et al. (2010) Ancient DNA from European early Neolithic farmers reveals their near eastern affinities. PLoS Biology 8, e1000536. Halbert ND (2003) The utilization of genetic markers to resolve modern management issues in historic Bison populations: implications for species conservation. PhD Texas A&M University. Halbert ND, Derr JN (2007) A comprehensive evaluation of cattle tntrogression into US federal Bison herds. Journal of Heredity 98, 1-12. Harney WE (1951) The dingo. Walkabout, 36-37. Hertwig ST, Schweizer M, Stepanow S, et al. (2009) Regionally high rates of hybridisation and introgression in German wildcat populations (Felis silvestris, Carnivora, Felidae). Journal of Zoological Systematics and Evolutionary Research 47, 283-297. Hewitt G (2000) The genetic legacy of the Quaternary ice ages. Nature 405, 907-913. Hewson-Freund HJ (2003) Inheritance of coat colour in the Australian Cattle Dog and the Australian Stumpy Tail Cattle Dog. In: A Dog Called Blue (ed. Clark NR), pp. 133-149. Clark, N. R. , Sydney. Hickerson MJ, Carstens BC, Cavender-Bares J, et al. (2010) Phylogeography’s past, present, and future: 10 years after. Molecular Phylogenetics and Evolution 54, 291-301.

182 Hindrikson M, Männil P, Ozolins J, Krzywinski A, Saarma U (2012) Bucking the trend in Wolf-dog hybridization: First evidence from Europe of hybridization between female dogs and male Wolves. PLoS ONE 7, e46465. Ho SYW, Lanfear R, Bromham L, et al. (2011) Time-dependent rates of molecular evolution. Molecular Ecology 20, 3087-3101. Ho SYW, Larson G (2006) Molecular clocks: when timesare a-changin'. Trends in genetics 22, 79-83. Hoelzel A, Dahlheim M, Stern S (1998) Low genetic variation among killer whales (Orcinus orca) in the eastern North Pacific and genetic differentiation between foraging specialists. Journal of Heredity 89, 121-128. Hong KW, Kim SW, Jang HC, et al. (2009) Melanocyte-stimulating hormone receptor (MC1R) genotype and its effects on coat color in Korean Jindo dogs. Asian- Australasian Journal of Animal Sciences 22, 1078-1084. Humphries EM, Winker K (2011) Discord reigns among nuclear, mitochondrial and phenotypic estimates of divergence in nine lineages of trans-Beringian birds. Molecular Ecology 20, 573-583. Huson HJ, Parker HG, Runstadler J, Ostrander EA (2010) A genetic dissection of breed composition and performance enhancement in the Alaskan sled dog. BMC Genetics 11, 71. Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23, 1801-1806. Johannes JE (2003) The Basenji Annual Estrus: The impact of the rainforest ecology on its development. The Basenji 39, 10-11. Johnson CN, Isaac JL, Fisher DO (2007) Rarity of a top predator triggers continent- wide collapse of mammal prey: dingoes and marsupials in Australia. Proceedings of Royal Society of London B: Biological Sciences 274, 341-346. Johnson CN, Ritchie EG (2012) The dingo and biodiversity conservation: response to Fleming et al. (2012). Australian Mammalogy 35, 8-14. Johnson CN, VanDerWal J (2009) Evidence that dingoes limit abundance of a mesopredator in eastern Australian forests. Journal of Applied Ecology 46, 641- 646. Johnson CN, Wroe S (2003) Causes of extinction of vertebrates during the Holocene of mainland Australia: arrival of the dingo, or human impact? The Holocene 13, 941-948. Jones E (2009) Hybridisation between the dingo, Canis lupus dingo, and the domestic dog, Canis lupus familiaris, in Victoria: a critical review. Australian Mammalogy 31, 1-7. Jones FW (1925) The mammals of South Australia. British Science Guild Handbooks Committee, Adelaide. Kaleski R (1933) Australian barkers and biters. Endeavour Press, Sydney. Kass RE, RafteryAE (1995) Bayes factors. Journal of the American Statistical Association 90, 773–795. Karafet TM, Lansing JS, Redd AJ, et al. (2005) Balinese Y-chromosome perspective on the peopling of Indonesia: genetic contributions from pre-neolithic hunter- gatherers, Austronesian farmers, and Indian traders. Human Biology 77, 93-114.

183 Kayser M, Brauer S, Weiss G, et al. (2001) Independent histories of Human Y chromosomes from Melanesia and Australia. The American Journal of Human Genetics 68, 173-190. Keller LF, Waller DM (2002) Inbreeding effects in wild populations. Trends in Ecology & Evolution 17, 230-241. Kennedy LJ, Angles JM, Barnes A, et al. (2007) DLA-DRB1, DQA1, and DQB1 alleles and haplotypes in North American Gray Wolves. Journal of Heredity 98, 491- 499. Kennedy M, Phillips BL, Legge S, Murphy SA, Faulkner RA (2012) Do dingoes suppress the activity of feral cats in northern Australia? Austral Ecology 37, 134-139. Kennedy PK, Kennedy ML, Clarkson PL, Liepins IS (1991) Genetic variability in natural populations of the gray wolf, Canis lupus. Canadian Journal of Zoology 69, 1183-1188. Kerns JA, Cargill EJ, Clark LA, et al. (2007) Linkage and segregation analysis of black and brindle coat color in domestic dogs. Genetics 176, 1679-1689. Kerns JA, Newton J, Berryere TG, et al. (2004) Characterization of the dog Agouti gene and a nonagouti mutation in German Shepherd dogs. Mammalian Genome 15, 798-808. Kerns JA, Olivier M, Lust G, Barsh GS (2003) Exclusion of Melanocortin-1 receptor (Mc1r) and Agouti as candidates for dominant black in dogs. Journal of Heredity 94, 75-79. Kidwell SM, Flessa KW (1996) The quality of the fossil record: Populations, species, and communities. Annual Review of Earth and Planetary Sciences 24, 433-464. Kingman, JFC (1982) The coalescent. Stochastic Processes and their Applications 13, 235-248. Kishino H, Hasegawa M (1989) Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. Journal of Molecular Evolution 29, 170-179. Klütsch CFC, Seppälä EH, Fall T, et al. (2011) Regional occurrence, high frequency but low diversity of mitochondrial DNA haplogroup d1 suggests a recent dog-wolf hybridisation in Scandinavia. Animal Genetics 42, 100-103. Kokko H, López-Sepulcre A (2006) From individual dispersal to species ranges: Perspectives for a changing world. Science 313, 789-791. Koler-Matznick J, Brisbin IL, Feinstein M, Bulmer S (2004) An updated description of the New Guinea singing dog (Canis hallstromi, Troughton 1957). Journal of Zoology 261, 109-118. Laikre L, Ryman N (1991) Inbreeding depression in a captive wolf (Canis lupus) population. Conservation biology 5, 33-40. Langergraber KE, Prüfer K, Rowney C, et al. (2012) Generation times in wild chimpanzees and gorillas suggest earlier divergence times in great ape and human evolution. Proceedings of the National Academy of Sciences 109, 15716- 15721. Larson G, Cucchi T, Fujita M, et al. (2007) Phylogeny and ancient DNA of Sus provides insights into neolithic expansion in Island Southeast Asia and Oceania. Proceedings of the National Academy of Sciences 104, 4834-4839.

184 Larson G, Karlsson EK, Perri A, et al. (2012) Rethinking dog domestication by integrating genetics, archeology, and biogeography. Proceedings of the National Academy of Sciences 109, 8878-8883. Larson G, Liu R, Zhao X, et al. (2010) Patterns of East Asian pig domestication, migration, and turnover revealed by modern and ancient DNA. Proceedings of the National Academy of Sciences 107, 7686-7691. Lee T-H, Guo H, Wang X, Kim C, Paterson AH (2014) SNPhylo: a pipeline to construct a phylogenetic tree from huge SNP data. BMC Genomics 15, 162. Leonard J, Wayne R, Wheeler J, et al. (2002) Ancient DNA evidence for Old World origin of New World dogs. Science 298, 1613. Leroy G, Verrier E, Meriaux J, Rognon X (2009) Genetic diversity of dog breeds: within‐ breed diversity comparing genealogical and molecular data. Animal Genetics 40, 323-332. Letnic M, Crowther MS (2012) Patterns in the abundance of kangaroo populations in arid Australia are consistent with the exploitation ecosystems hypothesis. Oikos 122, 761-769 Letnic M, Crowther MS, Dickman CR, Ritchie E (2011) Demonising the dingo: how much wild dogma is enough. Current Zoology 57, 668-670. Letnic M, Crowther MS, Koch F (2009a) Does a top-predator provide an endangered rodent with refuge from an invasive mesopredator? Animal Conservation 12, 302-312. Letnic M, Dworjanyn SA (2011) Does a top predator reduce the predatory impact of an invasive mesopredator on an endangered rodent? Ecography 34, 827-835. Letnic M, Fillios M, Crowther MS (2012a) Could direct killing by larger dingoes have caused the extinction of the Thylacine from mainland Australia? PLoS ONE 7, e34877. Letnic M, Greenville A, Denny E, et al. (2010) Does a top predator supress the abundance of an invasive mesopredator at a continental scale? Global Ecology and Biogeography 20, 343-353. Letnic M, Koch F (2010) Are dingoes a trophic regulator in arid Australia? A comparison of mammal communities on either side of the dingo fence. Austral Ecology 35, 167-175. Letnic M, Koch F, Gordon C, Crowther MS, Dickman CR (2009) Keystone effects of an alien top-predator stem extinctions of native mammals. Proceedings of Royal Society of London B: Biological Sciences 276, 3249-3256. Letnic M, Ritchie EG, Dickman CR (2012b) Top predators as biodiversity regulators: the dingo Canis lupus dingo as a case study. Biological Reviews 87, 390-413. Librado P, Rozas J (2009) DNAsp v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25, 1451-1452. Lindblad-Toh K, Wade CM, Mikkelsen TS, et al. (2005) Genome sequence, comparative analysis and haplotype structure of the domestic dog. Nature 438, 803-819. Little CC (1957) The inheritance of coat color in dogs. Howell Book House. Macintosh NW (1964) 4 thousand year old thylacine tooth (dasyuridae ) + 3 thousand year old dingo skeleton from D J Mulvaneys archaeological excavations at Fromms Landing Lower Murray River. Journal of Anatomy 98, 491-495.

185 Macintosh NWG (1975) The origin of the dingo: an enigma. In: The Wild Canids - their systematics, behavioural ecology and evolution (ed. Fox MW), pp. 87-106. Van Nostrand Reinhold Company, New York Marsack P, Campbell G (1990) Feeding-behavior and diet of dingoes in the Nullarbor Region, Western-Australia. Wildlife Research 17, 349-357. May SA, Norton TW (1996) Influence of fragmentation and disturbance on the potential impact of feral predators on native fauna in Australian forest ecosystems. Wildlife Research 23, 387-400. McEvoy BP, Lind JM, Wang ET, et al. (2010) Whole-Genome genetic diversity in a sample of Australians with deep Aboriginal ancestry. The American Journal of Human Genetics 87, 297-305. McNiven IJ (1998) Aboriginal archaeology of the Corroboree Beach dune field, Fraser Island: Re-survey and re-assessment. Memoirs of the Queensland Museum, Culture 1, 1. McRae BH, Beier P, Dewald LE, Huynh LY, Keim P (2005) Habitat barriers limit gene flow and illuminate historical events in a wide-ranging carnivore, the American puma. Molecular Ecology 14, 1965-1977. Miao YW, Wu GS, Wang L, et al. (2010) The role of MC1R gene in buffalo coat color. Science China Life Sciences 53, 267-272. Milham P, Thompson P (1976) Relative antiquity of Human occupation and extinct fauna at Madura Cave, Southeastern Western Australia. Mankind 10, 175-180. Millar SE, Millar MW, Stevens ME, Barsh GS (1995) Expression and transgenic studies of the mouse agouti gene provide insight into the mechanisms by which mammalian coat colour patterns are generated. Development 121, 3223-3232. Miller W, Drautz D, Ratan A, et al. (2008) Sequencing the nuclear genome of the extinct woolly mammoth. Nature 456, 387-390. Mitchell BD, Banks PB (2005) Do wild dogs exclude foxes? Evidence for competition from dietary and spatial overlaps. Austral Ecology 30, 581-591. Moseby KE, Neilly H, Read JL, Crisp HA (2012) Interactions between a top order predator and exotic mesopredators in the Australian rangelands. International Journal of Ecology 2012. Mulvaney DJ (1975) The prehistory of Australia. Penguin Books, Ringwood. Mulvaney DJ, Kamminga J (1999) Prehistory of Australia. Smithsonian Institution Press, Washington DC. Munoz-Fuentes V, Darimont C, Wayne R, Paquet P, Leonard J (2009) Ecological factors drive differentiation in wolves from British Columbia. Journal of biogeography 36, 1516 - 1531. Musiani M, Leonard J, Cluff H, et al. (2007) Differentiation of tundra/taiga and boreal coniferous forest wolves, coat color and association with migratory caribou. Molecular Ecology 16, 4149. Natanaelsson C, Oskarsson MC, Angleby H, et al. (2006) Dog Y chromosomal DNA sequence: identification, sequencing and SNP discovery. BMC Genetics 7, 45. Newsome A, Corbett L (1982) The identity of the dingo II. Hybridization with domestic dogs in captivity and in the wild. Australian Journal of Zoology 30, 365-374. Newsome AE, Corbett LK (1985) The identity of the dingo III. The incidence of dingoes, dogs and hybrids and their coat colours in remote and settled regions of Australia. Australian Journal of Zoology 33, 363-373.

186 Newsome AE, Corbett LK, Carpenter SM (1980) The identity of the dingo I. Morphological discriminants of dingo and dog skulls. Australian Journal of Zoology 28, 615-625. Newton JM, Wilkie AL, He L, et al. (2000) Melanocortin 1 receptor variation in the domestic dog. Mammalian Genome 11, 24-30. Norris BJ, Whan VA (2008) A gene duplication affecting expression of the ovine ASIP gene is responsible for white and black sheep. Genome Research 18, 1282-1293. Nowak RM (2003) Wolf evolution and taxonomy. In: Wolves: behavior, ecology, and conservation (eds. Mech LD, Boitani L), pp. 239–258. University of Chicago Press, Chicago (IL). O’Grady JJ, Brook BW, Reed DH, et al. (2006) Realistic levels of inbreeding depression strongly affect extinction risk in wild populations. Biological Conservation 133, 42-51. Ólafsdóttir GÁ, Kristjánsson T (2008) Correlated pedigree and molecular estimates of inbreeding and their ability to detect inbreeding depression in the Icelandic sheepdog, a recently bottlenecked population of domestic dogs. Conservation Genetics 9, 1639-1641. O'Neill A (2002) Living with the Dingo. Envirobook, Sydney. Oskarsson MCR, Klütsch CFC, Boonyaprakob U, et al. (2011) Mitochondrial DNA data indicate an introduction through Mainland Southeast Asia for Australian dingoes and Polynesian domestic dogs. Proceedings of the Royal Society B: Biological Sciences 279, 967-974 Pal SK, Ghosh B, Roy S (1998) Dispersal behaviour of free-ranging dogs (Canis familiaris) in relation to age, sex, season and dispersal distance. Applied Animal Behaviour Science 61, 123-132. Pang J-F, Kluetsch C, Zou X-J, et al. (2009) mtDNA data indicate a single origin for dogs south of Yangtze River, less than 16,300 years ago, from numerous Wolves. Molecular Biology and Evolution 26, 2849-2864. Parsons MH, Blumstein DT (2010) Familiarity breeds contempt: Kangaroos persistently avoid areas with experimentally deployed dingo scents. PLoS ONE 5, e10403. Philipp U, Hamann H, Mecklenburg L, et al. (2005) Polymorphisms within the canine MLPH gene are associated with dilute coat color in dogs. BMC Genetics 6. Pilot M, Jedrzejewski W, Branicki W, et al. (2006) Ecological factors influence population genetic structure of European grey wolves. Molecular Ecology 15, 4533-4553. Pople AR, Grigg GC, Cairns SC, Beard LA, Alexander P (2000) Trends in the numbers of red kangaroos and emus on either side of the South Australian dingo fence: evidence for predator regulation? Wildlife Research 27, 269-276. Pugach I, Delfin F, Gunnarsdóttir E, Kayser M, Stoneking M (2013) Genome-wide data substantiate Holocene gene flow from India to Australia. Proceedings of the National Academy of Sciences 110, 1803-1808. Pugach I, Stoneking M (2013) Reply to Price and Bird: No inconsistency between the date of gene flow from India and the Australian archaeological record. Proceedings of the National Academy of Sciences 110, E2949. Puja IK, Irion DN, Schaffer AL, Pedersen NC (2005) The Kintamani dog: Genetic profile of an emerging breed from Bali, Indonesia. Journal of Heredity 96, 854- 859.

187 Purcell S, Chang C (2014) PLINK 1.9. Accessed from https://www.cog- genomics.org/plink2 Purcell S, Neale B, Todd-Brown K, et al. (2007) PLINK: a tool set for whole-genome association and population-based linkage analyses. The American Journal of Human Genetics 81, 559-575. R Development Core Team (2010) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Räikkönen J, Vucetich JA, Vucetich LM, Peterson RO, Nelson MP (2013) What the inbred Scandinavian Wolf population tells us about the nature of conservation. PLoS ONE 8, e67218. Rambaut A, Drummond A (2007) Tracer Available from http://beast.bio.ed.ac.uk/Tracer. Randall D, Pollinger J, Argaw K, Macdonald D, Wayne R (2010) Fine-scale genetic structure in Ethiopian wolves imposed by sociality, migration, and population bottlenecks. Conservation Genetics 11, 89-101. Ripple WJ, Estes JA, Beschta RL, et al. (2014) Status and ecological effects of the world’s largest carnivores. Science 343, 1241484. Roberts RG (2014) A Pardon for the Dingo. Science 343, 142-143. Robertshaw J, Harden R (1986) The ecology of the dingo in Northeastern New-South- Wales .4. Prey selection by dingoes, and its effect on the major prey species, the Swamp Wallaby, Wallabia-Bicolor (Desmarest). Wildlife Research 13, 141-163. Robin S, Tacher S, Rimbault M, et al. (2009) Genetic diversity of canine olfactory receptors. BMC Genomics 10, 21 Robley A, Gormley A, Forsyth DM, Wilton AN, Stephens D (2010) Movements and habitat selection by wild dogs in eastern Victoria. Australian Mammalogy 32, 23-32. Rokas A, Williams BL, King N, Carroll SB (2003) Genome-scale approaches to resolving incongruence in molecular phylogenies. Nature 425, 798-804. Rosenberg NA (2004) DISTRUCT: a program for the graphical display of population structure. Molecular Ecology Notes 4, 137-138. Rothschild MF, Van Cleave PS, Glenn KL, Carlstrom LP, Ellinwood NM (2006) Association of MITF with white spotting in Beagle crosses and Newfoundland dogs. Animal Genetics 37, 606-607. Roughsey D (1973) The giant devil dingo. Collins, Sydney. Rubinoff D, Holland BS (2005) Between two extremes: Mitochondrial DNA is neither the panacea nor the nemesis of phylogenetic and taxonomic inference. Systematic Biology 54, 952-961. Rueness EK, Jorde PE, Hellborg L, et al. (2003a) Cryptic population structure in a large, mobile mammalian predator: the Scandinavian lynx. Molecular Ecology 12, 2623-2633. Rueness EK, Stenseth NC, O'Donoghue M, et al. (2003b) Ecological and genetic spatial structuring in the Canadian lynx. Nature 425, 69-72. Runstadler JA, Angles JM, Pedersen NC (2006) Dog leucocyte antigen class II diversity and relationships among indigenous dogs of the island nations of Indonesia (Bali), Australia and New Guinea. Tissue Antigens 68, 418-426. Sacks BN, Bannasch DL, Chomel BB, Ernest HB (2008) Coyotes demonstrate how habitat specialization by individuals of a generalist species can diversify

188 populations in a heterogeneous ecoregion. Molecular Biology and Evolution 25, 1384-1394. Sacks BN, Brown SK, Ernest HB (2004) Population structure of California coyotes corresponds to habitat-specific breaks and illuminates species history. Molecular Ecology 13, 1265-1275. Sacks BN, Brown SK, Stephens D, et al. (2013) Y chromosome analysis of dingoes and Southeast Asian village dogs suggests a Neolithic continental expansion from Southeast Asia followed by multiple Austronesian dispersals. Molecular Biology and Evolution 30, 1103-1118 Savolainen P, Leitner T, Wilton AN, Matisoo-Smith E, Lundeberg J (2004) A detailed picture of the origin of the Australian dingo, obtained from the study of mitochondrial DNA. Proceedings of the National Academy of Sciences 101, 12387-12390. Savolainen P, Zhang Y-p, Luo J, Lundeberg J, Leitner T (2002) Genetic evidence for an East Asian origin of domestic dogs. Science 298, 1610-1613. Schliep KP (2011) Phangorn: phylogenetic analysis in R. Bioinformatics 27, 592-593. Schmutz SM, Berryere TG (2007) Genes affecting coat colour and pattern in domestic dogs: a review. Animal Genetics 38, 539-549. Schmutz SM, Berryere TG, Barta JL, Reddick KD, Schmutz JK (2007) Agouti sequence polymorphisms in coyotes, wolves and dogs suggest hybridisation. Journal of Heredity 98, 351-355. Schmutz SM, Berryere TG, Dreger DL (2009) MITF and white spotting in dogs: A population study. Journal of Heredity 100, S66-S74. Schmutz SM, Berryere TG, Ellinwood NM, Kerns JA, Barsh GS (2003) MCIR studies in dogs with melanistic mask or brindle patterns. Journal of Heredity 94, 69-73. Shearin AL, Ostrander EA (2010) Canine morphology: Hunting for genes and tracking mutations. PLoS Biology 8, e1000310. Shepherd N (1981) Predation of Red Kangaroos, Macropus rufus, by the dingo, Canis familiaris dingo (Blumenbach) in North-Western New South Wales. Wildlife Research 8, 255-262. Shimodaira H, Hasegawa M (1999) Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16, 1114-1116. Shimodaira H, Hasegawa M (2001a) CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17, 1246–1247. Shimodaira, H (2001b) An approximately unbiased test of phylogenetic tree selection. Systematic Biology 51, 492–508. Shivik JA (2004) Non-lethal alternatives for predation management. Sheep & Goat Research Journal 19, 64-71 Shivik JA (2006) Tools for the edge: what's new for conserving carnivores. BioScience 56, 253-259. Silvestro D, Michalak I (2012). raxmlGUI: a graphical front-end for RAxML. Organisms Diversity and Evolution 12, 335-337 Siracusa LD (1994) The agouti gene: turned on to yellow. Trends in Genetics 10, 423- 428. Smith B, Litchfield C (2010) Dingoes (Canis dingo) can use human social cues to locate hidden food. Animal Cognition 13, 367-376.

189 Smith BP, Litchfield CA (2009) A review of the relationship between Indigenous Australians, dingoes (Canis dingo) and domestic dogs (Canis familiaris). Anthrozoös 22, 111-128. Stenglein JL, Waits LP, Ausband DE, Zager P, Mack CM (2011) Estimating gray wolf pack size and family relationships using noninvasive genetic sampling at rendezvous sites. Journal of Mammalogy 92, 784-795. Stenseth NC, Shabbar A, Chan K-S, et al. (2004) Snow conditions may create an invisible barrier for lynx. Proceedings of the National Academy of Sciences 101, 10632-10634. Stephens D (2011) The molecular ecology of Australian wild dogs: hybridisation, gene flow and genetic structure at multiple geographic scales PhD, University of Western Australia. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013) MEGA6: Molecular Evolutionary Genetics Analysis Version 6.0. Molecular Biology and Evolution 30, 2725-2729. Tajima F (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123, 585-595. Tappen M (1994) Bone weathering in the tropical rain forest. Journal of Archaeological Science 21, 667-673. Thalmann O, Shapiro B, Cui P, et al. (2013) Complete mitochondrial genomes of ancient canids suggest a European origin of domestic dogs. Science 342, 871- 874. Thomson P (1992) The behavioural ecology of dingoes in north-western Australia. II. Activity patterns, breeding season and pup rearing. Wildlife Research 19, 519- 529. Thomson P, Rose K, Kok N (1992) The behavioural ecology of dingoes in northwestern Australia. VI. Temporary extraterritorial movements and dispersal. Wildlife Research 19, 585-595. Toews DPL, Brelsford A (2012) The biogeography of mitochondrial and nuclear discordance in animals. Molecular Ecology 21, 3907-3930. Trigger D, Mulcock J, Gaynor A, Toussaint Y (2008) Ecological restoration, cultural preferences and the negotiation of ‘nativeness’ in Australia. Geoforum 39, 1273- 1283. Trinkel M, Cooper D, Packer C, Slotow R (2011) Inbreeding depression increases susceptibility to bovine tuberculosis in Lions: an experimental test using an inbred–outbred contrast through translocation. Journal of Wildlife Diseases 47, 494-500. van Bommel L, Johnson CN (2012) Good dog! Using livestock guardian dogs to protect livestock from predators in Australia’s extensive grazing systems. Wildlife Research 39, 220-229. van Holst Pellekaan S (2011) Genetic evidence for the colonization of Australia. Quaternary International. van Holst Pellekaan SM (2001) Origins of the Australian and New Guinean Aborigines. In: eLS. John Wiley & Sons, Ltd. Vilà C, Maldonado J, Wayne R (1999) Phylogenetic relationships, evolution, and genetic diversity of the domestic dog. Journal of Heredity 90, 71-77. Vila C, Savolainen P, Maldonado J, et al. (1997) Multiple and ancient origins of the domestic dog. Science 276, 1687 - 1689.

190 vonHoldt B, Pollinger J, Earl D, et al. (2012) Identification of recent hybridisation between gray wolves and domesticated dogs by SNP genotyping. Mammalian Genome, 1-9. vonHoldt BM, Pollinger JP, Earl DA, et al. (2011) A genome-wide perspective on the evolutionary history of enigmatic wolf-like canids. Genome Research 21, 1294- 1305. vonHoldt BM, Pollinger JP, Lohmueller KE, et al. (2010) Genome-wide SNP and haplotype analyses reveal a rich history underlying dog domestication. Nature 464, 898-902. vonHoldt BM, Stahler DR, Bangs EE, et al. (2010b) A novel assessment of population structure and gene flow in grey wolf populations of the Northern Rocky Mountains of the United States. Molecular Ecology 19, 4412-4427. Vrieling H, Duhl DMJ, Millar SE, Miller KA, Barsh GS (1994) Differences in dorsal and ventral pigmentation result from regional expression of the mouse Agouti gene. Proceedings of the National Academy of Sciences 91, 5667-5671. Wallach AD, Johnson CN, Ritchie EG, O’Neill AJ (2010) Predator control promotes invasive dominated ecological states. Ecology Letters 13, 1008-1018. Wallach AD, Murray BR, O’Neill AJ (2009a) Can threatened species survive where the top predator is absent? Biological Conservation 142, 43-52. Wallach AD, Ritchie EG, Read J, O'Neill AJ (2009) More than mere numbers: The impact of lethal control on the social stability of a top-order predator. PLoS ONE 4, e6861. Walters B (1995) The company of dingoes, two decades with our Native dog. Australian Native Dog Conservation Society Limited, Bargo. Wang Y, Fisher DO (2012) Dingoes affect activity of feral cats, but do not exclude them from the habitat of an endangered macropod. Wildlife Research 39, 611- 620. Ward TJ, Bielawski JP, Davis SK, Templeton JW, Derr JN (1999) Identification of domestic cattle hybrids in wild cattle and bison species: a general approach using mtDNA markers and the parametric bootstrap. Animal Conservation 2, 51- 57. Webb KM, Allard MW (2009) Mitochondrial genome DNA analysis of the domestic dog: Identifying informative SNPs outside of the control region. Journal of Forensic Sciences 54, 275-288. Weckworth BV, Dawson NG, Talbot SL, Flamme MJ, Cook JA (2011) Going coastal: Shared evolutionary history between coastal British Columbia and Southeast Alaska Wolves (Canis lupus). PLoS ONE 6, e19582. Wiens JJ (2004) Speciation and ecology revisited: Phylogenetic niche conservatism and the origin of species. Evolution 58, 193-197. Wiens JJ, Donoghue MJ (2004) Historical biogeography, ecology and species richness. Trends in Ecology & Evolution 19, 639-644. Wilton A (2001) DNA methods of assessing Australian dingo purity. In: A Symposium on the Australian dingo (eds. Dickman CR, Lunney D), pp. 49 - 55. Royal Zoological Society of New South Wales, Sydney. Wilton AN, Steward DJ, Zafiris K (1999) Microsatellite variation in the Australian dingo. Journal of Heredity 90, 108-111.

191 Wolff JN, Shearman DCA, Brooks RC, Ballard JWO (2012) Selective enrichment and sequencing of whole mitochondrial genomes in the presence of nuclear encoded mitochondrial pseudogenes (Numts). PLoS ONE 7, e37142. Woodall PF, Pavlov P, Twyford KL (1996) Dingoes in Queensland, Australia: skull dimensions and the indenity of wild canids. Wildlife Research 23, 581-587. Xu YC, Fang SG, Li ZK (2007) Sustainability of the South China tiger: implications of inbreeding depression and introgression. Conservation Genetics 8, 1199-1207. Zhang H-h, Wei Q-g, Zhang H-x, Chen L (2011) Comparison of the fraction of olfactory receptor pseudogenes in wolf (Canis lupus) with domestic dog (Canis familiaris). Journal of Forestry Research 22, 275-280. Zheng H-X, Yan S, Qin Z-D, et al. (2011) Major population expansion of East Asians began before Neolithic time: Evidence of mtDNA genomes. PLoS ONE 6, e25835. Zink RM, Barrowclough GF (2008) Mitochondrial DNA under siege in avian phylogeography. Molecular Ecology 17, 2107-2121.

192 Appendix

193 Appendix 1 List of co-authored publications

Cairns KM, Wilton AN & Ballard JWO (2011). The identification of dingoes in a background of hybrids. In: Advances in genetics research (ed. Urbano KV), pp.309–

327. Nova Science Publishers, New York.

194 Appendix 2

Table Appendix 2 Sample data; Identifier, Geographical Locale, Latitude, Longitude and Genetic Identity CR CR haplotype haplotype MtDNA MtDNA Y-Chr SNP Name ID Canid Sex State Longitude Latitude (collapsed) (gaps) lineage type Haplotype cluster Alpine 1 96.2 dingo M VIC -37.29 148.33 A209 A209 SE a9 H60-k11 SE Alpine 2 WD170 dingo M NSW -36.46 148.26 A29 A29 SE a3 H1-k1 SE Alpine 3 44.5 dingo M ACT -35.84 148.98 A29 A29 SE a3 H1-6t SE Alpine 4 135.10 dingo F VIC -37.07 148.58 A29 A179 SE a3 SE Alpine 5 119.1 dingo M VIC -36.17 147.99 A29 A179 NW d23 H60-k11 SE Fraser 1 21.3 dingo F QLD -25.25 153.17 A29 A179 SE f1 Fraser 2 21.1 dingo M QLD -25.17 153.28 A29 A179 SE f1 H60-n25 Fraser 3 184.4 dingo M QLD -25.18 153.28 A29 A179 SE f1 F I Fraser 4 184.1 dingo M QLD -25.70 153.03 A29 A179 SE f1 H60-n25 FI Fraser 5 21.4 dingo F QLD -25.79 153.08 A29 A179 SE f1 FI Fraser 6 21.2 dingo F QLD -24.92 153.28 A29 A179 SE a1 Fraser 7 184.2 dingo M QLD -25.73 153.03 - - - - - F I Gibson 1 9.40 dingo F WA -26.22 121.55 A203 A203 NW d5 NW Gibson 2 19.84 dingo F WA -27.30 123.06 A29 A29 NW d5 NW Gibson 3 DE17 dingo M WA -26.42 121.67 A200 A200 NW d5 H60-n24 NW Gibson 4 DE13 dingo F WA -25.45 122.90 A29 A29 NW d5 NW Gibson 5 18.35 dingo F WA -25.08 122.05 A200 A200 NW d5 NW Kimberley 1 3.45 dingo F WA -15.10 125.53 A29 A29 NW d5 NW Kimberley 2 7.27 dingo M WA -16.63 124.88 A9 A9 NW d7 H3-n4 NW Kimberley 3 19.11 dingo M WA -17.47 125.08 A9 A9 NW d7 H3-n20 NW Kimberley 4 4.55 dingo F WA -15.35 126.10 din27 din27 NW d5 NW 195

Table Appendix 2 Sample data; Identifier, Geographical Locale, Latitude, Longitude and Genetic Identity CR CR haplotype haplotype MtDNA MtDNA Y-Chr SNP Name ID Canid Sex State Longitude Latitude (collapsed) (gaps) lineage type Haplotype cluster Kimberley 5 24.96 dingo M WA -17.27 122.57 A29 A29 NW d5 H3-n4 Simpson 1 182.6 dingo F SA -26.65 140.35 A29 A179 NW d5 N W Simpson 2 X1777 dingo M NT 25.35 133.71 A200 din35 NW d5 H60-9k NW Simpson 3 142.3 dingo F SA -27.94 134.74 A200 A200 NW d4 Simpson 4 217.2 dingo F SA -26.91 134.06 A200 A200 NW d14 Simpson 5 X1783 dingo F NT -24.23 131.42 A200 A200 NW d5 N W Simpson 6 193.5 dingo M NT -26.09 140.41 A29 A29 NW d3 H60-n25 Captive 1 Gunyah dingo M Captive - - A29 A29 SE a3 H1-k7 Captive 2 N858 dingo M Captive - - A29 A205 NW d5 H60-k8 Captive 3 N895 dingo M Captive - - A29 A29 SE a3 H1-n7 Captive 4 N834 dingo M Captive - - A29 A29 SE a3 H3-12d Captive 5 N887 dingo M Captive - - A29 A198 NW d5 H60-n24 Central Australia 1 X167 dingo M QLD -28.06 143.98 din32 din32 NW d20 H1-6q Central Australia 10 TA95 dingo M NT -20.51 129.46 A29 A29 - - H60-n29 Central Australia 11 X2060 dingo M NT -16.79 137.53 A17 A17 - - H3-6z Central Australia 12 150.1 dingo M SA -34.81 139.25 din32 din32 NW d21 H60-k10 Central Australia 13 155.5 dingo M SA -31.76 136.35 din31 din31 NW d5 H60-n25 Central Australia 14 185.1 dingo M SA -29.38 138.03 A29 A179 NW d22 H1-6q Central Australia 4 197.1 dingo M SA -31.85 138.47 din32 din32 NW d5 H1-6q Central Australia 2 X179 dingo M QLD -28.06 143.98 din32 din32 NW d20 H1-6q Central Australia 3 182.2 dingo M NT -26.71 140.63 A29 A179 NW d5 H1-6q Central Australia 5 193.2 dingo F NT -26.72 135.07 A29 A29 NW d1 Central Australia 6 193.3 dingo F NT -26.71 135.08 - - NW d2

196 Central Australia 7 TA101 dingo M NT -20.49 129.32 A200 A200 NW d9 H60-0i

Table Appendix 2 Sample data; Identifier, Geographical Locale, Latitude, Longitude and Genetic Identity CR CR haplotype haplotype MtDNA MtDNA Y-Chr SNP Name ID Canid Sex State Longitude Latitude (collapsed) (gaps) lineage type Haplotype cluster Central Australia 8 TA91 dingo M NT -20.53 130.30 A200 A200 NW d9 H60-n24 Central Australia 9 TA94 dingo M NT -20.49 129.32 A29 A29 NW d5 H60-n25 Dubbo 146.1r dingo M NSW -32.15 148.85 A29 A199 NW d8 H3-k9 Inglewood X2654 dingo F QLD -28.41 151.08 - - NW d16 Moree X920 dingo M NSW -29.47 149.47 A29 A208 SE a10 H1-6t Northeastern 1 WD333 dingo F QLD -17.33 145.39 A29 A199 NW d16 Northeastern 10 157.1 dingo M QLD -20.74 144.01 A29 A199 NW d5 H60-9k Northeastern 11 219.2 dingo M QLD -16.40 145.36 A29 A199 NW d5 H60-n27 Northeastern 2 WD386 dingo M QLD -12.71 143.28 A29 A207 NW d16 H60-n28 Northeastern 3 WD402 dingo F QLD -20.06 146.27 A201 din34 NW d16 Northeastern 4 X2159 dingo M QLD -16.25 145.32 A29 A207 NW d16 H60-n27 Northeastern 5 X987 dingo F QLD -19.30 146.73 A29 A179 NW d16 Northeastern 6 X980 dingo F QLD -19.31 146.74 A29 A179 NW d16 Northeastern 7 X983 dingo F QLD -19.32 146.74 A29 A179 NW d16 Northeastern 8 X985 dingo M QLD -19.22 146.44 A29 A179 NW d16 H60-n24 Northwestern 1 1.64 dingo M WA -26.47 120.84 A200 din36 NW d5 H60-n17 Northwestern 10 18.38 dingo F WA -25.08 122.05 A29 A29 NW d12 Northwestern 11 3.47 dingo M WA -14.85 125.97 A29 A29 NW d5 H3-n4 Northwestern 12 4.52 dingo F WA -15.796 126.372 A29 A29 NW d5 Northwestern 13 7.36 dingo F WA -15.71 126.20 din27 din27 NW d18 Northwestern 14 21.23 dingo F WA -17.03 125.43 A29 A29 NW d5 Northwestern 15 DE11 dingo F WA -26.57 122.81 A200 A200 NW d5 Northwestern 16 14.95 dingo F WA -13.97 126.96 A29 A29 NW d24

197 Northwestern 17 3.48 dingo F WA -14.80 125.75 A29 A29 NW d5

Table Appendix 2 Sample data; Identifier, Geographical Locale, Latitude, Longitude and Genetic Identity CR CR haplotype haplotype MtDNA MtDNA Y-Chr SNP Name ID Canid Sex State Longitude Latitude (collapsed) (gaps) lineage type Haplotype cluster Northwestern 18 4.62 dingo M WA -15.33 126.42 A29 A29 NW d10 H3-n21 Northwestern 19 N866 dingo M WA -17.08 128.20 A202 A202 NW d11 H60-k11 Northwestern 2 11.55 dingo M WA -27.41 122.36 A29 A29 NW d15 H60-n22 N W Northwestern 20 X2581 dingo M WA -22.94 118.96 A210 din33 NW d5 H60-n25 Northwestern 21 X2583 dingo M WA -22.89 118.14 A203 A203 NW d5 H60-n22 Northwestern 22 X2601 dingo M WA -22.73 119.06 A29 A29 NW d5 H60-n25 Northwestern 23 X3291 dingo M WA -21.66 121.57 A29 A29 NW d5 H60-k3 Northwestern 3 3.46 dingo M WA -15.06 125.54 A29 A29 NW d5 H3-12d Northwestern 4 4.53 dingo M WA -15.42 126.14 A29 A29 NW d19 H3-n21 Northwestern 5 4.54 dingo F WA -15.70 126.36 A29 A29 NW d10 Northwestern 6 17.87 dingo M WA -16.81 125.71 A202 A202 NW d6 H3-n21 Northwestern 7 21.72 dingo F WA -22.33 122.08 A29 A179 NW d13 Northwestern 8 9.39 dingo F WA -25.73 125.78 A200 A200 NW d5 Northwestern 9 24.94 dingo F WA -17.27 122.57 A29 A29 NW d15 N W Southeastern 1 X1267 dingo M QLD -21.99 148.03 A213 A213 SE a10 H1-1c Southeastern 10 X874 dingo F NSW -35.75 148.26 A29 A29 SE a2 Southeastern 11 X931 dingo F NSW -35.85 148.21 A29 A29 SE a2 Southeastern 12 WD192 dingo F NSW -35.29 148.78 A29 A29 SE a2 Southeastern 13 X791 dingo F NSW -30.43 152.23 A29 A29 SE a11 Southeastern 14 44.2 dingo F ACT -35.80 148.91 A29 A29 SE a3 Southeastern 15 85.1 dingo F VIC -37.34 147.90 A29 A29 SE a3 Southeastern 17 21.5 dingo F QLD -25.45 153.07 A29 A179 SE a1 Southeastern 18 65.1 dingo F VIC -36.28 147.87 A29 A29 SE a2

198 Southeastern 19 166.4 dingo F VIC -36.44 147.97 A29 A29 SE a2

Table Appendix 2 Sample data; Identifier, Geographical Locale, Latitude, Longitude and Genetic Identity CR CR haplotype haplotype MtDNA MtDNA Y-Chr SNP Name ID Canid Sex State Longitude Latitude (collapsed) (gaps) lineage type Haplotype cluster Southeastern 2 X1273 dingo F QLD 27.07 152.97 A29 A179 SE a10 Southeastern 20 X1006 dingo M ACT -35.37 148.93 A29 A29 SE a3 H1-k1 Southeastern 21 X1012 dingo M ACT -35.89 149.04 A29 A29 SE a3 H1-6t Southeastern 22 X1049 dingo M ACT -35.43 148.88 A29 A29 SE a3 H1-k1 Southeastern 23 X1050 dingo M ACT -35.43 148.88 A29 A29 SE a3 H1-k1 Southeastern 24 X1062 dingo M ACT -35.36 148.92 A29 A29 SE a3 H1-k1 Southeastern 25 X2279 dingo M ACT -35.64 148.96 A29 A29 SE a3 H1-6t Southeastern 26 156.3 dingo M QLD -25.09 148.77 A29 A29 SE a7 H60-9k Southeastern 27 W0143 dingo M NSW -33.29 151.20 A29 A29 SE a3 H1-6t Southeastern 28 W0144 dingo M NSW -33.29 151.20 A29 A29 SE a3 H1-6t Southeastern 29 W0151 dingo M NSW -36.91 149.24 A29 A179 SE a3 H60-n24 Southeastern 3 X229 dingo F NSW -31.45 152.72 A29 A29 SE a8 Southeastern 30 WD036 dingo M NSW -35.78 148.01 A29 A29 SE a3 H3-k2 Southeastern 31 X1311 dingo M NSW -31.31 152.95 A29 A29 SE a1 H3-k9 Southeastern 32 X2256 dingo M NSW -35.90 149.05 A29 A29 SE a3 H1-k1 Southeastern 33 X2389 dingo M NSW -31.19 152.97 A29 A29 SE a1 H3-k9 Southeastern 34 X2405 dingo M NSW -31.00 153.02 A29 A29 SE a1 H3-k9 Southeastern 35 X2482 dingo M NSW -31.84 152.14 A29 A29 SE a6 H1-6t Southeastern 36 X2484 dingo M NSW -30.78 152.68 A29 A29 SE a1 H60-n25 Southeastern 37 X2529 dingo M NSW -30.43 152.23 A29 A29 SE a1 H1-k6 Southeastern 38 X2764 dingo M NSW -30.78 151.16 A29 A208 SE a5 H3-k9 Southeastern 39 X2792 dingo M NSW -32.55 152.31 A29 A29 SE a1 H60-k4 Southeastern 4 X1020 dingo M ACT -35.87 149.00 A29 A29 SE a2 H1-6t Southeastern 40 X2931 dingo M NSW -30.89 151.94 A29 A29 SE a1 H1-6t 199

Table Appendix 2 Sample data; Identifier, Geographical Locale, Latitude, Longitude and Genetic Identity CR CR haplotype haplotype MtDNA MtDNA Y-Chr SNP Name ID Canid Sex State Longitude Latitude (collapsed) (gaps) lineage type Haplotype cluster Southeastern 41 X3508 dingo M NSW -29.31 152.14 A29 A29 SE a1 H1-k6 Southeastern 42 X580 dingo M NSW -35.80 148.13 A29 A29 SE a3 H21-7d Southeastern 43 X606 dingo M NSW -35.88 148.66 A29 A29 SE a3 H60-k11 Southeastern 44 16.1 dingo M VIC -37.22 147.53 A29 A29 - - H3-k2 Southeastern 45 X2070 dingo M VIC -36.49 146.93 A29 A29 SE a4 H1-6t Southeastern 16 X296 dingo M QLD -25.48 153.06 A29 A179 SE f3 H60-n25 Southeastern 5 127.1 dingo F VIC -37.10 147.42 - - SE a4 Southeastern 6 96.4 dingo M VIC -37.22 148.16 A29 A29 SE a3 H1-k1 Southeastern 7 184.3 dingo M QLD -25.52 153.12 A29 A179 SE f2 H60-n25 Southeastern 8 144.8 dingo F NSW -35.78 148.25 A29 A29 SE a2 Southeastern 9 144.9 dingo F NSW -35.78 148.25 A29 A29 SE a2 NGSD 1 NGSD4 NGSD F PNG - - A79 A79 NGSD ng1 NGSD 2 NGSD2 NGSD F PNG - - A79 A79 NGSD ng1 NGSD 3 NGSD3 NGSD F PNG - - A79 A79 NGSD ng1 NGSD 4 NGSD6 NGSD M PNG - - A79 A79 NGSD ng1 H60-k10 NGSD 5 NGSD5 NGSD M PNG - - A79 A79 NGSD ng1 H60-k10 NGSD A Subu NGSD M PNG ------NG SD NGSD B Crosby NGSD M PNG ------NGSD NGSD C Keba NGSD M PNG ------NGSD NGSD D Hali NGSD F PNG ------NGSD NGSD E Roux NGSD M PNG ------NGSD Kalimantan X1764 dog F Indonesia 3.07 116.04 - - - - Bali X1772 dog F Indonesia -8.41 115.19 - - - -

200

Chapter 2 supplementary material

Appendix 3-1 PCR conditions and primers for Whole Mitochondrial Genome amplification and sequencing of the dingo and NGSD

Mitochondrial DNA PCR reactions were carried out in 25 µL volumes containing water, 5 X Crimson polymerase buffer (New England Biolabs Inc., MA, USA), 1.5 mM of MgCl2, 6.25 pmol of each primer pair (Table Appendix 3-1), 7.5 mM of dNTPs, 2.5

U of Taq DNA Polymerase (New England Biolabs Inc., MA, USA) and 20-50 ng of

DNA template.

All mitochondrial DNA PCR reactions were cycled using the following thermal profile:

98°C for 2 min, 95°C for 3 min (add Taq polymerase) then 95°C for 15 sec, 52°C for 1 min, 65°C for 1 min for 10 cycles, then 95°C for 15s, 52°C for 1 min, 65°C for 1 min

(increase time by 5 sec each cycle) for 25 cycles followed by 65°C for 10 min.

201

Table Appendix 3-1 Mitochondrial Primers for whole mitochondrial genome sequencing Primer Name Sequence Nucleotide position Reference G1and2_F GCTTAGCCCTAAACATAGAT 511 G1and2_R GTTTATGGGTAGCTCGTCT 1336 G3_F GTAGAGGTGAAAAGCCTAAC 1407 G3_R TGTTATGCCTTGTAGGTATC 2196 G4_F TTTGTTAGGGTGGCAGAG 2668 G4_R GATTCTCCTTCAGTCAAGTC 3342 G5_F TGATTACTGATCTGAATTGG 3990 G5_R GGTAGTCCTCCTAGGGATA 4660 G6_F TTCTACCTCACGAAATTTTA 5121 G6_R AGGCAGGGATAGTAAGAGTA 5930

G7_F TGAGCTTTAGGGTTTATTTT 6367 G7_R CATAATGGTTATGACATTGG 6993 Designed for this study by G8_F CCAATGATACTGAAGCTATG 7340 KMC

G8_R ATTTTAGCAGGTTTGGTTAT 7915 G9_F GAACCTGAGCTCTCATACTT 8161 G9_R TAAGACCCGATGTTATAAGG 8719 G10_F GTGTACGGATCTACCTTTTT 9218 G10_R TAGGAGGGAGATTAGTAGGA 9771 G11_F GCCCCTTATACTTATAGCAA 10422 G11_R CTTGTTATGATTATGCCTCA 11002 G12_F TCTCATATTTCCAAATAAAATC 13477 G12_R TAGTTGAAATACAACGATGA 14155 G13_F TGTACTACCATGAGGACAAA 14578 202

Table Appendix 3-1 Mitochondrial Primers for whole mitochondrial genome sequencing Primer Name Sequence Nucleotide position Reference G13_R CTAAAAGTCAGAATAGGCATT 15150 G14_F GCCCATGATCACACATAA 15880 G14_R AAAGCTTGTGAGTATTGTATG 16264 P1_F AAACCCCGATAAACCTCACC 638 P1_R GGTGTAGGCCAGACGCTTTA 1059 P2_F AAACAGCAAAGATTACCCCTTCTAC 1245 P2_R TGCTTTTAAGCCAACTATGGTGATA 1607 P3_F CTTGTATGAATGGCCACACG 2051

P3_R GTGAGGAAAGCTACGGCAAG 2808 P4_F CCTGACCCCTAGCCATGATA 3279 P4_R TATCATGATGGATGCGATGG 4207

P5_F GCTGAGGCGGACTAAATCAA 4410 P5_R CAAGCTCCGTGGTGAATTTT 5281 P6_F AGGAGCTCCGTTGACCTTA 5765

P6_R TCCTCCCATAATGGCAAAAA 6524 P7_F CTTACAGCGGTGATGCTTATAATTT 6732 P7_R GGTTATTTCTATTGGGAGGACAACT 7415

P8_F ATGGCCAGTGCTCTGAAATC 7611 Designed by Matthew Wong P8_R CGGTTAATCGAACGGCTAGA 8447 during his 2010 honours thesis (unpublished) and supervised P9_F TGCAGGACACCTCCTAATCC 8458 by author ANW. P9_R CGAAGTGGTGGTTTGATGTG 9343 P10_F CCCCTACGAATGTGGTTTTG 9600 P10_R ATGATCAGGGTCGGAATCAG 10595 P11_F ATACGGCCTCCACTTGTGAC 10827 P11_R TCTTATGCATGGGAGCATGA 11693 203

Table Appendix 3-1 Mitochondrial Primers for whole mitochondrial genome sequencing Primer Name Sequence Nucleotide position Reference P12_F ATGCCCTAATGACCTTGCAC 11501 P12_R GGCCGAATTGAGCAGATTTA 12462 P13_F CCCTTTTTGTCACGTGGTCT 11890 P13_R AAAATGCGTGAGTGCAGATG 12780 P14_F TTGGCATTAACCAGCCCTAC 12727 P14_R TATGGGACTTGGTTGGTGGT 13665 P15_F CCCGCTTCTCCCCTATAATC 13081 P15_R TATGGGACTTGGTTGGTGGT 13665 P16_F TTCAGAACAATCGCACAACC 13973 P16_R GGATCCGTTTCGTGTAGAA 14797 P17_F ATTCCATTTCACCCTTACTACACAA 14834 P17_R TATGCATTCGATTACTGTTGAAAGA 15660 P18_F ACTTCAGGGCCATAACCTTATTTAC 15798 P18_R ATGCGTATAAGACTGTTGTGTTAAG 16629 P19_F CACGCGCGTAAGACATTAAG 16431 P19_R GCGAAAGGTGGTGAGGTTTA 666 H15422 CTCTTGCTCCACCATCAGC 15422 Savolainen et al. 2004 L16106 AAACTATATGTCCTGAAACC 16106

204

Appendix 3-2 PCR conditions and primers for nuclear gene amplification and sequencing of the dingo and NGSD

All nuclear genes amplified using primers, thermal cycling and reaction conditions as described in original reference (Appendix 3-2 Table 1, Table 2 and Table 3).

205 Appendix 3-2 Table 1 Primers for DLA nuclear gene PCR amplification and sequencing. Gene Primer name Primer sequence Reference DQA1 DLA-DQAin1_F TAAGGTTCTTTTCTCCCTCT DLA-DQAin2_R GGACAGATTCAGTGAAGAGA DQB1 DLA-DQB1BT7_F TAATACGACTCACTATAGGGCTCACTGGCCCGGCTGTCTC Kennedy et al. DLA-DQBR2_R CACCTCGCCGCTGCAACGTG 2007 DRB1 DLA-DRB1_F GATCCCCCCGTCCCCACAG

DLA-DRB3_T7_R TAATACGACTCACTATAGGGCGCCCGCTGCGCTCA DQB1 and DRB1 T7 TAATACGACTCACTATAGGG sequencing Primer

206

Appendix 3-2 Table 2 Primers for Olfactory Receptor nuclear gene PCR amplification and sequencing. Gene Primer name Primer sequence Reference CfOR0007 CfOR0007_P_F TGCATTGCCATGAAATAAATATG Robin et al. 2009 CfOR0007_P_R ATCAGTGAATCTTGGCTGTAGGA CfOR0011 CfOR0011_P_F AAACCCGAGGAATACACAAAAAT Tacher et al. 2005 CfOR0011_P_R TGATGTTCATAGAACCCAAAAGG CfOR0034 CfOR0034_P_F TAATCCTAAGGTGGCACAAAAGA Robin et al. 2009 CfOR0034_P_R TGCTAATCTCAATGGGAATGTTT CfOR0123 CfOR0123_P_F ATTCCCCTGGTGAGGACTG Tacher et al. 2005 CfOR0123_P_R TGGCAACCTCCATATCTTCC CfOR0184 CfOR0184_P_F TCAAGTTCACCTTACCTGCAAA CfOR0184_P_R AAGGATCCCTGATTACCATGT

CfOR14A11 CfOR14A11_P_F ATTTGGGGGACTGACCATACTT Robin et al. 2009

CfOR14A11_P_R AATTTTGGTTTACATGTCTGGAG

DOPRH07 DOPRH07_P_F CTCGCTCTCTGGCAAATAAATAA

DOPRH07_P_R GGGTTAAGTGAGGTCTTCCAGAT

207

Appendix 3-2 Table 3 Primers for Coat Colour nuclear gene PCR amplification and sequencing. Gene Primer name Primer sequence Reference CBD103 CBD103-wolf-F TGTCTTCATCCCTGTGAGGT CBD103-wolf-R CCAGGAGGCATTTTCACACT ASIP ASIP-gDNA-4F AAGTCCAGCGGACAGTCG Anderson et al. ASIP-gDNA-4R CACACCTTGGAGCAGCCTA 2009 MC1R MC1Ramp1F CACTTGTACAGACCGGGAGAG

MC1Ramp2R AAATGCCCAGCAGGATAGTG MC1Ramp2F GTGACGAATGTGCTGGAGAC MC1Ramp3R ATCCACCACACCACAGATCA

208

209