Genomic and Breeding Resources to Produce Seeded and High Biomass Interspecific Hybrids of Napiergrass and Pearl Millet

GENOMIC AND BREEDING RESOURCES TO PRODUCE SEEDED AND HIGH BIOMASS INTERSPECIFIC HYBRIDS OF NAPIERGRASS AND PEARL MILLET

DEV RAJ PAUDEL

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2018

To my late Mom

ACKNOWLEDGMENTS

I wish to express my appreciation to the members of my advisory committee: Dr.

Fredy Altpeter, Dr. Jianping Wang, Dr. Patricio Munoz, Dr. Calvin Odero, and Dr.

Salvador Gezan, who have guided, supported, and encouraged me throughout the course of my research project. Sincere thanks to my research advisor Dr. Fredy

Altpeter for allowing me to join his group and pursue the work described here. Thank you for all the support, guidance, and mentorship that you have provided during my graduate studies. I am truly inspired by your professionalism and leadership role.

I am particularly indebted to my co-advisor Dr. Jianping Wang who provided me with space in her lab to do my experiments and provided me with opportunities to develop skills in molecular and computational biology. Your mentorship has been an invaluable gift over the past couple of years. One day, I hope to inspire others as you've inspired me.

I would like to gratefully acknowledge the University of Florida Graduate School

Fellowship for funding the first four years of my PhD. I am thankful to the Graduate

School Doctoral Dissertation Award for funding my final semester. I am grateful to the

Florida Plant Breeders Working Group for providing funds for this research.

I was grateful to be surrounded by many wonderful people during my journey.

Many thanks to talented undergraduates, Fan Wen, Erik Hanson, and Stephanie Maya, who helped not only with the experiments, but also helped me learn and hone my mentorship skills.

In addition, I would like to express my gratitude to the following people:

• Dr. Baskaran Kannan without whose help, support, and wisdom, this work

wouldn’t have been completed. 4

• Staff at Plant Science Research & Education Unit, Citra, FL for coordinating and

helping in planting, managing, and harvesting of research plots.

• Dr. Calvin Odero, Venancio Fernandez, Raphael Mereb Negrisoli, and Nikol

Havranek at the Everglades Research and Education Center for providing

support to do fieldwork in Belle Glade, FL.

• Dr. Tina Strauss, Dr. Eshan Gurung, Er. Prasan Gurung, and Dr. Laxman

Adhikari for your continued support, valuable insights in my research, and

friendship.

• Members of the Altpeter Lab: Derek Hurley, Dr. Saroj Parajuli, Dr. Ratna Karan,

Dr. Simon Gere, and Bryant Brown, for helping during field work and research

activities.

• Members of the Wang Lab: Dr. Liping Wang, Dr. Xiping Yang, Dr. Ze Peng, Dr.

Zhou Hai, Aleksey Kurashev, Dr. Yu-Chien Tseng, and Dr. Song Jian for helping

in fieldwork, lab experiments, as well as discussions related to research that

were instrumental in providing insights into some of the work described in this

document.

• Marco Sinche for kindly providing yield and flowering data of mapping

experiments.

• Dr. Rajeev Varshney for providing the pearl millet reference genome sequence.

• Dr. Karen-Harris Schultz for providing napiergrass sequences.

• Cynthia Hight for playing a phenomenal role in supporting academic endeavors

and ensuring that I was on track every semester.

• Members of the Plant Science Council for guidance, support, friendship, and

humor that has helped me survive graduate school.

• Nepalese community in Gainesville, FL for always being there for me.

Obtaining this advanced degree was only possible due to the sheer sacrifice, patience, unconditional love and understanding, long-term support, and encouragement in all aspects of my career and life from the love of my life, my dear wife, Ashmita

Guragain, my father Bala Bhadra Paudel, mother late Dhan Kumari Paudel, brother

Dipendra Paudel, sister-in-law Kalpana Sapkota, and niece Deleena Paudel. I am thankful to my parents-in-law and family, and other relatives for persistent encouragement and support.

Finally, this dissertation is dedicated to my true champion and idol, my mother, late Dhan Kumari Paudel, whom I lost during the final semester. She was my confidant and mentor. She was the epitome of love, sacrifice, simplicity, wisdom, and strength, to whom I owe everything. Life isn't the same without you.

I hope I have made you proud.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 10

LIST OF FIGURES ...... 12

LIST OF ABBREVIATIONS ...... 17

ABSTRACT ...... 18

CHAPTER

1 BACKGROUND ...... 20

Introduction ...... 20 Botany ...... 21 Napiergrass Ancestry ...... 22 Napiergrass in the United States ...... 23 Pearl Millet ...... 24 Interspecific Hybrids of Napiergrass and Pearl Millet (PMN Hybrids) ...... 25 Cytoplasmic Male Sterility (cms) ...... 25 Molecular Tools Applied in Plant Breeding ...... 27 Genotyping by Sequencing (GBS) ...... 29 Target Enrichment Sequencing ...... 30 Quantitative Trait Loci Analysis ...... 30 Biomass yield ...... 31 Flowering Time ...... 31 Genes Related to Flowering ...... 33 Objectives ...... 34

2 SURVEYING THE GENOME AND CONSTRUCTING A HIGH-DENSITY GENETIC MAP OF NAPIERGRASS (CENCHRUS PURPUREUS SCHUMACH.) ...... 37

Introduction ...... 38 Methods ...... 41 Napiergrass Genome Survey ...... 41 SSR Identification and Marker Development ...... 42 Plant Materials and DNA Extraction ...... 43 Genotyping-by-sequencing ...... 43 Comparative Genomics ...... 43 Sequence Analysis and SNP Calling ...... 44 Linkage Map Construction ...... 45 Comparison Between Napiergrass and Pearl Millet Genome...... 46 7

Results ...... 46 Napiergrass Genome Survey ...... 46 Genotyping-by-sequencing ...... 47 SNP Calling by Various SNP Callers ...... 49 Genetic Linkage Map Construction ...... 49 Comparison Between Genomes of Napiergrass and Pearl Millet ...... 51 Discussion ...... 52

3 MAPPING QTLS CONTROLLING FLOWER NUMBER AND FLOWERING TIME IN NAPIERGRASS ...... 84

Introduction ...... 84 Materials and Methods...... 87 Development of a Mapping Population ...... 87 Phenotyping the Mapping Population ...... 88 Genetic Map ...... 88 QTL Analysis ...... 89 Candidate Gene Identification ...... 90 Results ...... 91 Phenotypes ...... 91 Number of flowers ...... 91 Flowering Time ...... 91 QTL Analysis ...... 91 Number of flowers ...... 91 Flowering time ...... 92 Candidate Genes ...... 92 Discussion ...... 93 Conclusion ...... 97

4 EVALUATE THE GENETIC BACKGROUND OF FLOWERING TIME IN A NAPIERGRASS GERMPLASM COLLECTION ...... 113

Introduction ...... 113 Materials and Methods...... 117 Plant Materials and Phenotyping ...... 117 DNA Extraction ...... 117 Targeted Candidate Genes ...... 118 Probe Design ...... 118 Probe Synthesis, Selection, and Sequencing ...... 119 Sequence Read Trimming and Mapping ...... 120 SNP Calling ...... 120 Population Structure ...... 120 Results ...... 121 Candidate Genes Related to Flowering ...... 121 Probe Design ...... 121 Sequence processing ...... 123 Phylogenetic analysis ...... 124

Genome wide association analysis ...... 124 Discussion ...... 125 Conclusion ...... 129

5 GENERATION OF INTERSPECIFIC HYBRIDS BETWEEN PEARL MILLET AND NAPIERGRASS AND EVALUATION OF THEIR PERFORMANCE...... 142

Introduction ...... 142 Materials and Methods...... 147 Description of Male Sterile Line of Pearl Millet ...... 147 Production of cms Lines of Pearl Millet ...... 147 Production of PMN Hybrids ...... 148 Experimental Design ...... 149 Traits Evaluated ...... 149 Data Analysis ...... 150 Results ...... 150 Plant Height ...... 150 Tiller Number ...... 150 Stem Diameter ...... 151 Leaf Length ...... 151 Leaf Width ...... 152 Plant Biomass ...... 152 Dry Biomass ...... 153 Coefficient of Variation ...... 153 Correlation ...... 154 Discussion ...... 154

6 CONCLUDING REMARKS ...... 179

Summary ...... 179 Future work ...... 181

LIST OF REFERENCES ...... 183

BIOGRAPHICAL SKETCH ...... 208

LIST OF TABLES

Table page

2-1 Parameters used for SNP calling for each software ...... 58

2-2 Repetitive elements present in the napiergrass genome ...... 61

2-3 The sequence alignment of ten napiergrass sequence contigs to the pearl millet genome ...... 62

2-4 Frequency of classified repeat types (considering sequence complementary) in napiergrass ...... 63

2-5 Primer pairs developed for napiergrass SSR markers ...... 67

2-6 Summary of the alignment of non-redundant tags of napiergrass (Cenchrus purpureus) to the available genomes of different species ...... 68

2-7 Alignment of individual napiergrass reads using Bowtie2 ...... 69

2-8 Summary of the combined linkage map of napiergrass and the percentage of gaps less than 5 cM in male and female parent linkage maps ...... 70

2-9 Summary of napiergrass single nucleotide polymorphism (SNP) markers mapped on the combined linkage map using 9 different software pipelines ...... 71

3-1 Descriptive statistics of flowering date and number of flowers for 185 F1 hybrids of a cross (N190  N122) at Citra, FL., in 2012, 2013 (Sinche 2013) and 2016 ...... 109

3-2 Detailed information about the QTLs for number of flowers. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM...... 110

3-3 Detailed information about the QTLs for flowering time. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM...... 111

3-4 List of putative flowering related genes from the genome of pearl millet (Varshney et al. 2017)...... 112

4-1 Genes and publications related to flowering used in this research ...... 130

5-1 Details of the cross types used in the experiment...... 175

5-2 Descriptive statistics for different types of crosses. P-value represents Pr>F for Levene’s test for homogeneity of variance (center = median), CV = coefficient of variation...... 176

5-3 Correlation coefficients and p-values for biomass weight and biomass-related traits for PMN hybrids evaluated in Citra, FL...... 177

LIST OF FIGURES

Figure page

1-1 Distribution of napiergrass. Black dots represent local napiergrass and red dots represent napiergrass listed as invasive species ...... 36

2-1 Sequence variation for SNPs called in various regions of the pearl millet genome...... 72

2-2 Micro-collinearity between contigs from napiergrass to the pearl millet genome...... 73

2-3 Inversion duplication between napiergrass and pearl millet (shown in bottom figure). 74

2-4 Estimated coverage of PstI restriction sites in the pearl millet genome...... 75

2-5 Histogram of uniquely mapped reads to the pearl millet genome...... 76

2-6 Venn diagram showing concordant napiergrass SNPs called by five reference-based SNP callers, SAMtools, GBS-SNP-CROP, GATK, FreeBayes, and TASSEL. Numbers in parenthesis after the program name shows the total number of SNPs called by each program...... 77

2-7 Genetic linkage map of the napiergrass female parent N190...... 78

2-8 Genetic linkage map of the napiergrass male parent N122...... 79

2-9 Genotyping by sequencing single nucleotide polymorphism (GBS-SNP) marker distribution for the 14 linkage groups of napiergrass. A black bar means a GBS-SNP marker. A blue bar represents segregation distortion region. The left scale plate represents genetic distance (centiMorgan as unit). . 80

2-10 Consensus genetic linkage map of napiergrass...... 81

2-11 Circos plot of the mapped TASSEL de-novo UNEAK napiergrass markers with pearl millet reference genome. Pearl millet pseudomolecules start with “PM” and are color coded for each pseudomolecule. Napiergrass linkage groups start with “LG” and are in green color. Each line corresponds to tags that showed significant BLAST hits to the pearl millet genome (> 80% identity and > 50 bp length)...... 82

2-12 Syntenic regions between napiergrass linkage groups and the pearl millet genome. PM01 to PM07 are pearl millet pseudomolecules, LG01 to LG14 are napiergrass linkage groups. The small dots represent significant BLAST hits of mapped UNEAK tags to the pearl millet genome (>80% identity and >50 bp length)...... 83

3-1 Histogram of number of flowers in the mapping population in 2012 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013...... 98

3-2 Histogram of number of flowers in mapping population for 2013 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013...... 99

3-3 Scatterplot of number of flowers between 2012 and 2013 in Citra, FL. X-axis represents the number of flowers in 2012 and y-axis represents the number of flowers in 2013. Data is adapted from Sinche, 2013...... 100

3-4 Histogram of number of days to first flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013...... 101

3-5 Histogram of number of days to first flowering in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013...... 102

3-6 Histogram of first flowering date in the mapping population in 2016 in EREC, Belle Glade, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted...... 103

3-7 Scatterplot of first date of flowering between different years and locations. X- axis and Y-axis represent the number of days to flowering (FT) in different years. Data for 2012 and 2013 are from Citra, FL and adapted from Sinche, 2013. Data for 2016 are from Belle Glade, FL...... 104

3-8 Scatterplot between number of flowers and days to flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2012 and y- axis represents the number of flowers per line in 2012. Data is adapted from Sinche, 2013...... 105

3-9 Scatterplot between days to flowering and number of flowers in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2013 and y- axis represents the number of flowers per line in 2013. Data is adapted from Sinche, 2013...... 106

3-10 Linkage map of male parent N122 showing potential and stable QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers

are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group...... 107

3-11 Linkage map of female parent N190 showing potential QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group...... 108

4-1 Histogram of length of flowering related genes...... 131

4-2 Number of probes designed per gene...... 132

4-3 Number of probes designed as a factor of the size of the gene...... 133

4-4 Distribution of the targeted flowering genes in the genome of pearl millet. Each blue bar represents a flowering gene mapped on the pearl millet genome...... 134

4-5 Number of paired-end reads per sample...... 135

4-6 Bayesian Information Criteria (BIC) vs. number of clusters in k-means clustering suggests K=3 in the germplasm collection...... 136

4-7 Projection of the napiergrass germplasm collection using first two linear discriminants (LDs) from discriminant analysis of principal components (DAPC). The shape of the points represent grouping by DAPC (circle = group 1; triangle = group 2; square = group 3) and the colors represent the origin continent: black = Americas; blue = Asia; and cyan = Africa...... 137

4-8 Evolutionary relationships of taxa. G stands for group assigned by DAPC and are in different shapes (circle = group 1; triangle = group 2; square = group 3) and the fill colors represent the origin: black = Americas, blue = Asia, cyan = Africa, and white=unknown)...... 138

4-9 Histogram for days to flowering trait in napiergrass germplasm collection...... 139

4-10 QQ plots for different models using GWASpoly using 78,129 SNP markers. DTF = Days to flowering...... 140

4-11 Manhattan plots for different models using GWASpoly. P values adjusted with FDR at 0.05. DTF = Days to flowering...... 141

5-1 Seeds of pearl millet dwarf cms line (A), cms high biomass pearl millet line (B), PMN hybrid (C), and napiergrass (D), respectively from left to right...... 158

5-2 Boxplot of plant height (cm) for four different types of crosses studied. Small colored dots represent individual plant height of the cross. Different letters in

blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 159

5-3 Histogram of plant height (cm) for four different crosses. X-axis represent plant height in cm and y-axis represents the count of plants. A = Tift 85 × 25- 17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 160

5-4 Boxplot of number of tillers for four different types of crosses studied. Small colored dots represent number of tillers for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25- 17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 161

5-5 Histogram of number of tillers for four different crosses. X-axis represent the number of tillers and y-axis represents the count of plants. A = Tift 85 × 25- 17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 162

5-6 Boxplot of stem diameter (mm) for four different types of crosses studied. Small colored dots represent the stem diameter (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787- S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 163

5-7 Histogram of stem diameter (mm) for four different crosses. X-axis represent the stem diameter (mm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17. ... 164

5-8 Boxplot of leaf length (cm) for four different types of crosses studied. Small colored dots represent the leaf length (cm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25- 17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 165

5-9 Histogram of leaf length (cm) for four different crosses. X-axis represent the leaf length (cm) and y-axis represents the count of plants. A = Tift 85 × 25- 17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 166

5-10 Boxplot of leaf width (mm) for four different types of crosses studied. Small colored dots represent the leaf width (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25- 17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 167

5-11 Histogram of leaf width (mm) for four different crosses. X-axis represent the leaf width (mm) and y-axis represents the count of plants. A = Tift 85 × 25- 17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 168

5-12 Boxplot of fresh biomass per plant (kg) for four different types of crosses studied. Small colored dots represent the fresh biomass per plant (kg) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25- 17...... 169

5-13 Histogram of fresh biomass per plant (kg) for four different crosses. X-axis represent the fresh biomass per plant (kg) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 170

5-14 Boxplot of projected dry biomass (tons per ha) for four different types of crosses evaluated at seven months of growth. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 171

5-15 Histogram of dry biomass (tons per ha.) for four different crosses evaluated at seven months of growth. X-axis represent the dry biomass (tons per ha.) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 172

5-16 Coefficient of variation (%) for the different traits evaluated. A = Tift 85 × 25- 17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17...... 173

5-17 Picture of four different crosses during harvest in Citra, FL. Cross A = Tift 85 × 25-17, Cross B = 787-S5 × 25-17, Cross C = MS 787 BC4 × 25-17, Cross D = P787 × 25-17...... 174

LIST OF ABBREVIATIONS cms Cytoplasmic male sterility

GBS Genotyping by sequencing

LG Linkage Group

NGS Next-generation sequencing

PMN Pearl millet napiergrass

SNP Single nucleotide polymorphism

SSR Simple Sequence Repeat

TES Targeted exome sequencing

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

GENOMIC AND BREEDING RESOURCES TO PRODUCE SEEDED AND HIGH BIOMASS INTERSPECIFIC HYBRIDS OF NAPIERGRASS AND PEARL MILLET

Dev Raj Paudel

December 2018

Chair: Fredy Altpeter Cochair: Jianping Wang Major: Agronomy

Napiergrass (Cenchrus purpureus Schumach) is a promising candidate for forage and lignocellulosic biofuel production due to its high biomass potential. However, napiergrass is listed as an invasive species in Florida due to wind dispersed seeds.

Seed production in napiergrass is compromised by low temperatures. Therefore, late flowering genotypes, which are not able to flower before low temperatures come in

Florida, could be utilized to improve biosafety of napiergrass. However, genetic and genomic resources for napiergrass are limited that preclude exploiting marker assisted selection (MAS) for crop improvement. Genetic linkage map is an important tool for

MAS. In this research, we constructed the first high-density genetic map of napiergrass by genotyping-by-sequencing a bi-parental mapping population of 185 F1 hybrids. As a result, we mapped 1,913 single nucleotide polymorphism (SNP) markers into 14 linkage groups of napiergrass, spanning a length of 1,410 cM with a density of one marker per

0.73 cM.

This genetic map enabled us to identify three stable and three potential quantitative trait loci (QTLs) controlling number of flowers and flowering time,

respectively in the mapping population. We also identified five candidate genes related to flowering in close proximity to the QTLs detected.

Full characterization of germplasm collections is very critical to efficiently utilize them in breeding programs. We used targeted enrichment sequencing to characterize the napiergrass germplasm collection and identified 78k SNPs in the collection. We inferred the structure of the germplasm collection and constructed its phylogeny.

Genome wide association studies revealed one significant SNP for flowering time.

Napiergrass is mostly vegetatively propagated, which makes the planting process much complicated and labor intensive. To ameliorate this, we introgressed cytoplasmic male sterility (cms) into elite lines of pearl millet and hybridized them with napiergrass to produce seed derived, sterile pearl millet napiergrass (PMN) hybrids. We evaluated the biomass yield and uniformity of PMN hybrids generated by using different parental backgrounds. There was a tremendous variation in different biomass related traits among the crosses. These seeded-yet-sterile PMN hybrids could have a major impact in the forage and biofuel industry if large scale production of high-quality seeds that give rise to high yielding progenies can be developed.

CHAPTER 1 BACKGROUND

Introduction

Napiergrass (Cenchrus purpureus Schumach, syn. Pennisetum purpureum

Schumach) is a tropical perennial grass that originated from Africa (Singh, Singh, and

Obeng 2013). It is commonly called merker grass, elephant grass, or Uganda grass.

Napiergrass is a monocotyledonous flowering plant that belongs to the grass family

(Poaceae) and genus Cenchrus. Recently, species of Pennisetum and Odontelytrum were merged and designated to the unified genus Cenchrus (Chemisquy et al. 2010).

Pennisetum genus consists of a heterogeneous group of over 140 species (Brunken

1977) and it includes species with basic chromosome number of 5, 7, 8 or 9 and ploidy ranging from diploid to octoploid (Martel et al. 1997).

Napiergrass is an important fodder crop as well as an important cellulosic energy crop due to its high dry biomass yield compared to sorghum, maize, sugarcane, switchgrass, johnsongrass, and Erianthus (Ra et al. 2012). The first published note on this grass was by Mynhardt, a Hungarian missionary in Barume (then north west

Rhodesia), who sent material to Zurich Botanical Gardens in 1905. However, the name napiergrass recognizes the contribution of Colonel Napier of Bulawayo who first wrote to the Rhodesian Agricultural Department to make them aware of the value of this plant

(Boonman 1988). Napiergrass is a short-day plant and flowering in tropical climates occurs from autumn through winter (Singh, Singh, and Obeng 2013). It is an open- pollinated species. The cultivars were mainly developed by vegetative propagation of superior clones derived from natural out-crossing material (Bhandari, Sukanya, and

Ramesh 2006; Augustin and Tcacenco 1993). Germplasm collections of napiergrass

are maintained in several countries in Africa, Brazil, Puerto Rico, United States,

Australia, China, Pakistan, and India (Azevedo et al. 2012; Bhandari, Sukanya, and

Ramesh 2006).

Botany

Napiergrass is a perennial, rhizomatous grass, propagated vegetatively, commonly from joints or cuttings of the canes (Thompson 1919). It tillers vigorously, and mature tillers have 20 or more internodes with plant height ranging from 2 to 6 m.

Its leaves are long and narrow. Reproductive part of this grass is panicles, characterized by tawny or purplish color, sessile fascicles, and sparsely plumose bristles (Singh, Singh, and Obeng 2013). Napiergrass is protogynous where the stigma exerts 3-4 days, starting at the top of the inflorescence, before anthesis, followed by anthers shedding pollen for 3-4 days. The seeds are tiny in size averaging 3.8 million seeds/kg and seed set is poor and prone to shattering with low germination rate.

Napiergrass requires a day length of 11 h or less to flower (Hanna 2004). When it is grown in sub-temperate regions, it produces no or very few flowers because by the time days shorten for flowering, the temperature drops too low to support flower growth.

Deep roots and rhizomes of napiergrass bind up the soil and prevent soil erosion (Juma

2014). The plant can be harvested in the first year itself unlike other grasses like

Miscanthus, that require the first year for establishment.

Napiergrass adapts to a wide range of soil types and pH range, while maximum growth is attained on well-drained loamy soils that have high organic matter. It is susceptible to waterlogging and traps cereal stem borer insect pest. Cereal farmers in western Kenya control pests like stem borer using push-pull technology by establishing a hedge of napiergrass around their cereal plots (Khan et al. 2001). 21

Napiergrass is highly heterozygous and gives rise to a very heterogeneous population of seedlings, that are not morphologically uniform (Juma 2014). High morphological variability of napiergrass acts as a valuable source of genetic variation.

Mass selection and hybridization has been used to develop several varieties of napiergrass such as Uganda hairless, Cameroon, Gold Coast, and Clone 13 (Juma

2014).

Napiergrass Ancestry

Napiergrass is an allotetraploid (2n=4x=28, A’A’BB) (Jauhar 1981) with two sub- genomes, A’ and B. The chromosomes in the A’ sub-genome of napiergrass is homologous to the A genome of pearl millet (Cenchrus americanus (L.) Morrone, syn.

Pennisetum glaucum, 2n=2x=14) (Jauhar 1981). The two species, napiergrass and pearl millet, can naturally hybridize to produce hybrids that are triploids with AA’ B genomes, thus are sterile (Singh, Singh, and Obeng 2013). Based on chromosome pairing in triploid hybrids (2n=3x=21; AA’B), it was inferred that the two species basically share a genome (A and A’ being very similar). Genomic in situ hybridization revealed a high level of homeology between genomes A and A’ (Reis et al. 2014). Approximately

29% of napiergrass genomes (A’B) were hybridized by the genome of pearl millet. It was inferred that napiergrass and pearl millet had concomitantly diverged from the common ancestor. The origin of napiergrass occurred at the interspecific hybridization event, combining genome A from the common ancestor with genome B whose source is still unknown (Reis et al. 2014; Jauhar and Hanna 1998). Genome A is 24% larger than genome A’ of napiergrass. This difference in genome size could be related to genic duplication in pearl millet and to genomic rearrangements in napiergrass. Pearl millet has genome DNA content of 4.72 pg while the genome DNA content of napiergrass is 22

4.60 pg (Martel et al. 1997). Both species have approximately similar DNA content (pg) but are different with regards to the basic chromosome number. Napiergrass is a tetraploid and has about half the DNA content (1.15 pg) of the pearl millet monoploid genome (2.36pg). Napiergrass also has smaller chromosomes than that of pearl millet

(Martel et al. 1997; Reis et al. 2014).

Napiergrass in the United States

Napiergrass was first introduced to Florida by forage breeders based on its high- yielding forage attributes and now occurs on the banks of canals, waterways and roadsides. Napiergrass cultivars that flower early produce a large amount of wind dispersed seeds that enhance its invasive potential. Napiergrass can spread to natural vegetation (D’Antonio and Vitousek 1992; Loope, Hamann, and Stone 1988; Schofield

1989) (Figure 1-1) and is listed as an invasive species by The Florida Exotic Plant Pest

Council (FLEPPC 2011). Therefore, environmental biosafety has become one of the major concerns in napiergrass cultivation. It is also a major concern for sugarcane cultivation as both have similar growth habits and reproduction modes (Singh, Singh, and Obeng 2013) making it difficult to manage in the crop. In order to enhance the environmental biosafety of napiergrass, measures to prevent invasiveness need to be implemented. Napiergrass usually flowers as day lengths decreases and late flowering has been observed in a number of accessions. Seed formation in napiergrass is compromised due to low temperatures at the end of the season. To improve the biosafety of napiergrass, it is important to select plants for late flowering and hybridization. Therefore, the development of cultivars that are sterile or are late flowering is important for napiergrass management in the field to mitigate invasiveness.

Pearl Millet

Pearl millet is a highly cross-pollinated monocot (Rajaram et al. 2013) belonging to Paniceae tribe within the Panicoideae subfamily of the Poaceae family (Devos and

Gale 2000; Devos et al. 2000). It is a coarse grass grown primarily for grain in Africa and Asia and as a feed and forage in America (Gupta and Mhere 1997). Pearl millet accounts for almost half of the global millet production and is important for food security in some of the world’s driest and hottest areas. It is grown mostly for its ability to produce grain under hot, dry conditions on infertile soils with low water-holding capacity

(ICRISAT and FAO 1996; Rajaram et al. 2013). Pearl millet has much larger seeds compared to napiergrass and is easy to establish in the field via seeds. Availability of a reference draft genome of pearl millet (Varshney et al. 2017) provides an excellent resource that can be used extensively for genomic improvement of pearl millet and related species.

Cytogenetic studies on napiergrass and pearl millet have classified pearl millet as primary and napiergrass as the secondary gene pool of the genus Pennisetum (Harlan and Wet 1971; Martel et al. 1997). Primary gene pool of the genus Pennisetum includes three species (one cultivated and two wild species) with 2n=2x=14 chromosomes.

Secondary gene pool is represented by one allotetraploid species, Pennisetum purpureum (2n=4x=28) with A’A’BB genomes (Martel, Ricroch, and Sarr 1996). The primary gene pool corresponds to the traditional concept of biological species and includes spontaneous races (wild and/or weedy) as well as cultivated races. Crossing and gene transfer is generally easy among species in the primary gene pool to form generally fertile hybrids. Species in the secondary gene pool can hybridize with species in the primary gene pool, but the hybrids tend to be sterile (Harlan and Wet 1971). 24

Interspecific Hybrids of Napiergrass and Pearl Millet (PMN Hybrids)

The hybrids of pearl millet and napiergrass, commonly called PMN hybrids, combine the superior forage quality of pearl millet and the high yielding ability of napiergrass (Gupta and Mhere 1997; Osgood, Hanna, and Tew 1997). These hybrids are male and female sterile due to triploidy (2n=3x=21) (Gupta and Mhere 1997) and have reduced persistence and ratooning ability as compared to napiergrass (Cuomo,

Blouin, and Beatty 1996). The resulting hybrid with AA’B genome is morphologically more similar to napiergrass due to larger genetic contribution (66.7% chromosomes) and dominance of napiergrass grass B genome over the pearl millet A genome for genetic characters such as earliness, inflorescence and leaf characteristics, and seed size (Obok, Ova, and Iwo 2012; Gonzalez and Hanna 1984). The dominance of B genome over the A’ genome masks genetic variability (consequently phenotypic variability) on the A’ genome. Most of the characteristics like resistance to pests, vigorous growth, and outstanding forage yield potential are contributed by the B genome (Hanna 1987). PMN hybrids do not set seed and thus do not pose a threat of uncontrolled establishment through dissemination of seeds (Hanna and Monson 1980) and are not considered invasive (Jessup 2013). Vegetative propagation of PMN hybrids or napiergrass is a major cost factor. Development of varieties that allow large scale production of high quality seeds giving rise to sterile plants represents a significant step in biomass grass breeding because field establishment using seeds will allow for automation resulting in significant cost reduction (Osgood, Hanna, and Tew 1997).

Cytoplasmic Male Sterility (cms)

Cytoplasmic male sterility (cms) results from the interaction between organellar and nuclear genomes that renders functional pollen production obsolete (Ram, Ram,

and Yadava 2007). cms plants do not produce viable pollen while being completely female fertile. cms is important in hybrid seed production, because it helps to control pollination for commercial production of F1 hybrid seeds (Smith and Chowdhury 1983). cms has been extensively and cost effectively exploited in a number of agronomic and horticultural crops including pearl millet (Havey 2004). cms in pearl millet was discovered in the 1955 (Burton 1958) and was first released as male-sterile inbred

Tifton 23A (A1 or milo cytoplasm). Other cms sources were also later identified (Burton and Athwal 1967) and the A4 cytoplasm was found useful for forage hybrids that do not require male-fertility restoration (Havey 2004). The availability of the cms pearl millet

‘Tift 23A’ paved the way to commercially produce seed-propagated interspecific hybrids that can facilitate seed harvest and allows the production of hybrids from easily drilled seeds (J. B. Powell and Burton 1966). cms can be transferred to elite varieties to make them sterile. To facilitate this, the method of backcrossing is suitable, where established cultivars that are deficient in one or two specific traits can be improved through crossing. In a backcross method, F1 hybrid is repeatedly crossed with a desirable parent to get the desired traits (Schlegel 2009; Acquaah 2012). This has been commonly used to transfer entire sets of chromosomes from foreign cytoplasm in order to create cytoplasmic male-sterile genotypes (Acquaah 2012). The availability of cms lines of pearl millet will help to develop homogenous lines of pearl millet that are male sterile.

These lines can then be used to cross with near-inbred napiergrass to produce uniform progenies of male and female sterile PMN hybrids. While producing seeds at a commercial scale, it should be noted that napiergrass usually flowers in November –

December, so seed production should be established in areas that stay frost-free until

December.

Molecular Tools Applied in Plant Breeding

As a group of important forage and biofuel species, napiergrass has received increasing attention in recent years. Molecular studies in napiergrass have concentrated on assessing genetic diversity, cultivar identification, origin, and relatedness using limited genetic marker technologies (López et al. 2014; Harris-Shultz, Anderson, and

Malik 2010; Kandel et al. 2016; Dowling, Burson, and Jessup 2014; Dowling et al.

2013). Despite some advances in genetic research of napiergrass, the genetic and molecular mechanisms of agronomic traits that must be improved for commercialization are still poorly understood. A good genetic understanding of biomass related traits such as emergence date, flowering time control, nutrient uptake, abiotic and biotic stress tolerance is needed to aid genetic improvement of crops which can be facilitated by a genetic linkage map (Ma et al. 2012) that helps to understand genome structure and to identify trait-specific molecular markers (Cai et al. 2015). In order to accelerate molecular tool development high-throughput genotyping is required (F. Lu et al. 2013).

Molecular markers have recently gained popularity because they are not subject to environmental influence, are available in vast numbers, and are more objective as compared to phenotypic markers (Kim et al. 2015). Today, plant breeders have access to several types of DNA markers and molecular breeding tools to choose from.

A DNA marker is a specific DNA sequence on a chromosome that shows polymorphism between individuals (Andersen and Lübberstedt 2003; Agarwal,

Shrivastava, and Padh 2008; Kumar 1999). Genomic variation leading to DNA polymorphisms can be directly linked to differences in phenotype, used to find 27

relationships between individuals in a population, and can be used as genetic markers

(Deschamps, Llaca, and May 2012; Rafalski 2002). Most crop agronomic traits are controlled by several loci that contribute to minor phenotypic effects and are known as quantitative trait loci (QTL)s (Falconer, Mackay, and Frankham 1996). Since, molecular markers are stable and detectable in all tissues and are not confounded by the environment and other effects (Agarwal, Shrivastava, and Padh 2008), it is possible to assign chromosomal positions to individual QTLs and to determine the effect of individual QTLs (Kumar 1999). In order to utilize these markers in a breeding program, these markers should be tightly linked to the desired QTL. One of the major objectives of molecular breeding in plant species is to connect phenotype to the genotype and then use this knowledge in MAS by making phenotypic predictions based on genotypes

(Poland and Rife 2012). Various marker technologies are currently available like restriction fragment length polymorphism (RFLP), random amplified polymorphic DNA

(RAPD), amplified fragment length polymorphism (AFLP), simple sequence repeat

(SSR), and single nucleotide polymorphism (SNP) (Agarwal, Shrivastava, and Padh

2008; Kumar 1999).

SSRs or microsatellites are short (2-6bp) repetitive DNA sequence distribute randomly in genomes. Since SSR markers are hyper-variable, multi-allelic, often codominant, highly reproducible, and readily multiplexed, they are one of the best choices for foreground selection in marker-assisted programs (Rajaram et al. 2013). They have been commonly used in genotyping populations, constructing linkage maps, and identifying QTLs for desirable traits. Linkage maps create a framework for trait mapping and QTL analysis (Rajaram et al. 2013). With the advent of next generation sequencing

(NGS) platforms, genotyping a large number of progenies of mapping populations become much efficient, which further speedup the QTL identification.

In recent years, SNP markers have gained much consideration in the breeding community as they occur in abundance and every SNP in a single copy DNA is a potentially useful marker (Ganal, Altmann, and Röder 2009). SNPs can be categorized according to nucleotide substitutions either as transitions (C/T or G/A) or transversions

(C/G, A/T, C/A or T/G) (Jiang 2013). SNPs are typically identified from sequences, such as expressed sequence tag (EST) sequence data, array analysis, amplicon sequencing, and next generation sequencing technologies (Ganal, Altmann, and Röder 2009). For

SNP genotyping, the widely used platforms include BeadXpressTM, GoldenGateTM,

Infinitum®, GeneChipTM, GenFlex TM, SNAPshotTM, TaqMan TM, SNPstreamTM,

SNPWaveTM, iPLEX GoldTM, ARRAYTM, and KASPTM among others. (Semagn et al.

2014).

Genotyping by Sequencing (GBS)

Sequence-based genotyping methods like GBS have enabled simultaneous marker discovery and genotyping (Poland and Rife 2012). GBS uses direct genome sequencing to produce genotyping information utilizing high throughput and multiplexing capacities of NGS (Paux et al. 2012; Deschamps, Llaca, and May 2012). Restriction enzyme is utilized for complexity reduction followed by multiplex sequencing, which requires less DNA as well as avoids random shearing and size selection (Poland,

Brown, et al. 2012). The resulting sequences allow direct SNP detection, which avoids the marker assay development stage (Deschamps, Llaca, and May 2012) and thus make GBS an efficient genotyping method (Elshire et al. 2011). GBS has proven to be effective for marker discovery and trait mapping in several species like wheat (Poland, 29

Endelman, et al. 2012), switchgrass (F. Lu et al. 2013), and potato (Uitdewilligen et al.

2013). Since, GBS performs SNP discovery and genotyping simultaneously, it is particularly of high value to understudied crops lacking reference genomes (Kim et al.

2015).

Target Enrichment Sequencing

Target Enrichment Sequencing (TES) is another NGS-enabled approach, which focuses on genes or genomic regions of interest for sequencing. In this approach, probes are designed to enrich capturing of the target genomic regions based on sequence homology, which are then sequenced. TES has been used in several crops such as barley (Russell et al. 2016), switchgrass (Grabowski et al. 2016), and peanuts

(Peng et al. 2017), to identify sequence variations but hasn't been previously applied in napiergrass.

Quantitative Trait Loci Analysis

Quantitative trait loci (QTL) refers to a specific region on chromosomes that harbors gene(s) controlling traits. QTL analysis is performed by estimating correlation between phenotype data with genotype (markers) data of segregating populations

(Miles and Wayne 2008). Primary types of segregating population for QTL mapping include F2, recombinant inbred lines (RILs), BC1 (backcross 1), double haploid lines

(DHL), near-isogenic lines (NILs) and full-sib F1 (pseudo-testcross) (Schneider 2005).

Quantitative traits, such as yield, height, flowering time, pest and disease resistance, in plants have been mapped in genomes of several species. QTLs are valuable information to develop markers linked to traits of interests for MAS in molecular breeding programs. Genetic variability for flowering time exists in napiergrass (Sinche

2013). Therefore, identifying flowering time related QTLs will help in identifying markers

linked to late flowering trait, an important trait to reduce invasiveness. Sexual hybridization of genetically distant parents and selection of late flowering, high yielding accessions would increase the biofuel yield and enhance the biosafety of napiergrass.

Biomass yield

In a comparison of major energy crops for ethanol production, napiergrass showed the highest dry biomass yield than sorghum, maize, sugarcane, switchgrass, johnsongrass, and Erianthus (Ra et al. 2012). Napiergrass yields up to 84.8 Mg ha-1 yr-1 have been obtained in Puerto Rico (Vicente-Chandler, Silva, and Figarella 1959), while in Florida, yields ranged between 35-45 Mg ha-1 yr-1 (Woodard and Sollenberger 2012;

Erickson et al. 2012). Biomass yield is a complex trait and studies have been conducted to identify QTLs for various components affecting yield. In sorghum, QTLs for yield related traits such as plant height, tiller number, leaf length, leaf width, stem diameter, and flowering time have been identified (Hart et al. 2001; Murray et al. 2008; Xiao-ping et al. 2011). Similarly, QTLs for leaf yield, stem yield, plant height and flowering time have also been identified in Miscanthus sinensis (Gifford et al. 2015; Atienza et al.

2003). In Miscanthus, QTLs for yield co-segregated with other traits like number of tillers, leaf area, leaf length, and leaf width (Gifford et al. 2015). In napiergrass, biomass yield showed high correlations with number of tillers, plant height, and stem diameter

(Sinche 2013).

Flowering Time

Floral transition is the switch from vegetative growth to reproductive growth in plants and it primarily determines flowering time. Flowering time is a key factor in plant adaptation and is linked to various attributes like plant height, yield, number of leaves, etc. (Durand et al. 2012). In many species, flowering is induced in response to the

length of periods of lightness and darkness associated with day and night length. These species are categorized as short-day, long-day, intermediate-day, or day-neutral based on their day length requirement (Schlegel 2009; Bastow and Dean 2002). Plants in which flowering is favored by day lengths shorter than the critical and corresponding long nights are called short-day plants (eg. Glycine max, Oryza sativa, and Zea mays).

The plants in which flowering is initiated when the day length is longer than the critical are called long-day plants (eg. Hordeum vulgare, Triticum aestivum, and Solanum tuberosum) (Garner 1933; Schlegel 2009; Bastow and Dean 2002). Napiergrass belongs to the short-day plants (Osgood, Hanna, and Tew 1997; Singh, Singh, and

Obeng 2013).

Many studies have detected QTLs related to flowering time or earliness in various crops. The genetic control of flowering time is in general quantitative in nature

(H. Lu et al. 2014). For example, in rice, 15 QTLs were associated with days to flowering (Maheswaran et al. 2000), and in tomato, three QTLs related to earliness were identified that were associated with flowering time, fruit set time, and ripening time

(Lindhout et al. 1994). However, a previous study in maize showed that flowering time was a complex trait and no major QTL should be expected (Buckler et al. 2009). QTLs for flowering related traits like 50% anthesis and heading date colocalized with other

QTLs for number of tillers, tiller diameter, leaf width, and leaf area in Miscanthus

(Gifford et al. 2015). In orchardgrass (Dactylis glomerata L)., 11 QTLs for heading date and flowering time were found to be distributed on three linkage groups where candidate genes such as hd1 and VRN1 were annotated (Zhao et al. 2016).

Flowering time in napiergrass is of prime importance as it is related to biosafety and biomass quality. Late flowering napiergrass lines serve as potential bio-safe biofuels as their flowering is compromised by low temperatures which usually occurs in early December in Florida. Late flowering cultivars, therefore have less potential for invasiveness than early flowering genotypes and may produce higher yields due to a longer period of vegetative growth. Extensive phenotypic variation in flowering time is an indication that flowering time is quantitative in nature and several genes might be controlling the trait. Identifying QTLs for flowering time in napiergrass will help to shorten the breeding cycle by the successful utilization of MAS. A biparental mapping population from a cross between early-flowering and late-flowering lines will generate progenies segregating for flowering time and this population can be genotyped in order to identify QTLs for flowering time.

Genes Related to Flowering

Flowering time has been studied in some grass species and genomic studies have identified a number of genes involved in flowering. Potential candidates can be

FRI, LEAFY, CO, DNF, MADS-box, and RID1 that have some roles in flowering regulation in other species like Arabidopsis, wheat, maize, rice (Lee, Bleecker, and

Amasino 1993; Wuxing Li et al. 2013; Suárez-López et al. 2001; Morris et al. 2010; C.

Wu et al. 2008). For example, high levels of proteins encoded by FLOWERING LOCUS

T (FT) are correlated with early flowering and the lack of these causes late flowering

(Samach 2012). Major genes involved in photoperiod of flowering are highly conserved between rice and Arabidopsis (C. Wu et al. 2008). For example, Hd1 QTL in rice that promotes heading under short-day conditions corresponds to a gene homolog of

CONSTANS in Arabidopsis (Yano et al. 2000); Hd17 corresponds to a homolog of 33

Arabidopsis ELF3 (EARLY FLOWERING 3) (Matsubara et al. 2012; Matsubara et al.

2008); and Hd3a encodes a protein that is closely related to Arabidopsis FT (Kojima et al. 2002). Mining of genome sequences that are available for several grass species for flowering related genes and their characterization can identify candidate genes in napiergrass. These candidate genes can be used for screening germplasm collections for identification of haplotypes that confer late flowering. In addition to this, integrating genomics with conventional breeding will help to shorten the breeding cycle for selection.

Objectives

The overall objectives of this research are to identify genetic components (QTL and candidate genes) underlying flowering time in napiergrass and to develop male and female sterile PMN hybrids that are yielding high amounts of biomass and display uniform seed progenies.

The substantial variation in flowering time in napiergrass can be utilized for breeding late flowering varieties. This helps to make napiergrass a bio safe genotype for biomass production in northern Florida where a freeze event typically occurs before flowering of these late lines. The identification of QTLs controlling flowering time will make the pre-selection for late flowering lines in breeding materials highly efficient.

Identifying polymorphisms within flowering genes in napiergrass germplasm collection helps to reveal allelic variation in the germplasm. Development of a male sterile and homozygous line of pearl millet will prevent its self-pollination, thus facilitating easy crossing with napiergrass under field conditions. These high-biomass yielding male sterile lines can then be used for hybridization with napiergrass.

The specific objectives of this research were to 1) construct genetic maps of napiergrass, 2) identify QTLs for flowering time in napiergrass, 3) evaluate sequence variations in a napiergrass germplasm collection, and 4) develop and evaluate male and female sterile PMN hybrids.

0 Latitude

-50

-100 0 100 200 Longitude

Figure 1-1. Distribution of napiergrass. Black dots represent local napiergrass and red dots represent napiergrass listed as invasive species

CHAPTER 2 SURVEYING THE GENOME AND CONSTRUCTING A HIGH-DENSITY GENETIC MAP OF NAPIERGRASS (CENCHRUS PURPUREUS SCHUMACH.)

Napiergrass (Cenchrus purpureus Schumach.) is a tropical forage grass and a promising lignocellulosic biofuel feedstock due to its high biomass yield, persistence, and nutritive value. However, its utilization for breeding has lagged behind other crops due to limited genetic and genomic resources. In this study, next-generation sequencing was first used to survey the genome of napiergrass. Napiergrass sequences displayed high synteny to the pearl millet genome and showed expansions in the pearl millet genome along with genomic rearrangements between the two genomes. An average repeat content of 27.5% was observed in napiergrass including 5,339 simple sequence repeats (SSRs). Furthermore, to construct a high-density genetic map of napiergrass, genotyping-by-sequencing (GBS) was employed in a bi-parental population of 185 F1 hybrids. A total of 512 million high quality reads were generated and 287,093 SNPs were called by using multiple de-novo and reference-based SNP callers. Single dose

SNPs were used to construct the first high-density linkage map that resulted in 1,913

SNPs mapped to 14 linkage groups, spanning a length of 1,410 cM and a density of 1 marker per 0.73 cM. This map can be used for many further genetic and genomic studies in napiergrass and related species.

This chapter was published in Scientific Reports and is licensed under a Creative Commons Attribution 4.0 International License. Paudel, D., Kannan, B., Yang, X., Harris-Shultz, K., Thudi, M., Varshney, R.K., Altpeter, F. and Wang, J., 2018. Surveying the genome and constructing a high-density genetic map of napiergrass (Cenchrus purpureus Schumach). Scientific reports, 8(1), p.14419.

Introduction

Napiergrass (Cenchrus purpureus Schumach., syn. Pennisetum purpureum

Schumach.), also known as elephant grass, is a tropical perennial grass native to eastern and central Africa. It is cultivated primarily for forage and widely used by smallholder dairy farmers due to its high growth rate, leaf nutritive value, perennial nature, persistence, ease of propagation, and broad adaptation (Bhandari, Sukanya, and Ramesh 2006; Farrell, Simons, and Hillocks 2002; Singh, Singh, and Obeng 2013;

Chemisquy et al. 2010). As a C4 grass species, napiergrass is a promising candidate feedstock for biofuel production due to its superior yield of biomass (Ra et al. 2012;

Anderson, Casler, and Baldwin 2008; Somerville et al. 2010). Napiergrass cultivars are typically developed from natural out-crossings (Bhandari, Sukanya, and Ramesh 2006;

Augustin and Tcacenco 1993). It is an allotetraploid (2n = 4x = 28, A’A’BB) (Jauhar

1981) with an average amount of DNA per G1 nucleus of 5.78 pg (M. G. Taylor and

Vasil 1987). The chromosomes in the A’ genome of napiergrass are believed to be homologous to the A genome of pearl millet (Pennisetum glaucum, 2n = 2x = 14, AA)

(Jauhar 1981). Pearl millet and napiergrass form a monophyletic group (Martel et al.

1997) and were initially classified as primary and secondary gene pool of the genus

Pennisetum, respectively (Harlan and Wet 1971; Martel et al. 1997). Recently, species of Pennisetum and Odontelytrum were transferred to the unified genus Cenchrus

(Chemisquy et al. 2010). Pearl millet and napiergrass can hybridize to produce hybrids called kinggrass (Dowling, Burson, and Jessup 2014) or Pearl Millet-Napiergrass (PMN) hybrids (Hanna and Monson 1980; Burton 1944; Muldoon and Pearson 1979). These hybrids are sterile due to triploidy (2n = 3x = 21, AA’B genome) (Gupta and Mhere

1997), thus preventing the unintended spreading into natural areas or other cropping 38

systems by wind dispersed seeds. Some PMN hybrids show high heterosis for biomass yield and forage quality while the perennial, persistent nature is often reduced compared to napiergrass (Singh, Singh, and Obeng 2013).

The targeted improvement of napiergrass includes identification of agronomically superior genotypes and studies assessing genetic diversity and relatedness using random amplification of polymorphic DNA (RAPD), amplified fragment length polymorphism (AFLP), isozymes, and simple sequence repeats (SSRs) developed for other species like pearl millet and buffelgrass (Pennisetum ciliare) (Lowe et al. 2003;

Bhandari, Sukanya, and Ramesh 2006; Harris-Shultz, Anderson, and Malik 2010;

Kandel et al. 2016; Dowling et al. 2013; Dowling, Burson, and Jessup 2014; López et al.

2014; Smith et al. 1993). Other than these, genetic information on napiergrass is very meager (Smith et al. 1993). A genetic map is lacking and molecular tools are not yet deployed in napiergrass breeding programs (Dowling et al. 2013; Negawo et al. 2017).

Development of molecular markers for detection and utilization of DNA polymorphisms will help to understand the molecular basis of various agronomic traits (Song et al.

2015). Molecular breeding for yield components, flowering date, nutrient uptake, abiotic and biotic stress tolerance will accelerate genetic improvement of napiergrass. This can be greatly facilitated by having access to marker resources like SSR, single nucleotide polymorphisms (SNPs), and genetic linkage maps. SSRs as molecular markers are very advantageous because they are locus specific, multi-allelic, co-dominant, and easy to detect by polymerase chain reaction (PCR) (W. Powell, Machray, and Provan 1996;

Kannan et al. 2014). SNP markers have gained increasing consideration in molecular breeding and linkage map construction as they occur in a large number and high

density (Ganal, Altmann, and Röder 2009). Access to these resources will support marker-assisted selection (MAS) by making phenotypic predictions based on the genotype (Poland and Rife 2012).

Recently, next generation sequencing (NGS) technology has simplified linkage map construction by using high throughput genotyping-by-sequencing (GBS), which allows simultaneous SNP discovery and genotyping across the whole genome of the population of interest (Elshire et al. 2011; Poland and Rife 2012; Deschamps, Llaca, and May 2012). GBS has been effective for marker discovery, genetic mapping, quantitative trait locus (QTL) analysis, population genetics, and comparative genomics studies in several diploid species, and has recently gained popularity in polyploid species such as wheat (Triticum aestivum) (Poland, Endelman, et al. 2012), switchgrass

(Panicum virgatum) (F. Lu et al. 2013), potato (Solanum tuberosum) (Uitdewilligen et al.

2013), and sugarcane (Saccharum spp.) (Yang, Sood, et al. 2017) among others.

However, the presence of highly similar homeologous copies of two genomes in allopolyploid species complicates SNP detection which relies on delineating true allelic

SNPs from homeologous SNPs because sequences from homeologous loci mimic allelic SNPs and can introduce false-positives. Distinguishing allelic SNPs from homeologous SNPs relies on the use of high-stringency sequence read alignment, specifically uniquely aligned reads (Clevenger et al. 2015). Despite challenges of using

GBS for genotyping of polyploid species, genetic mapping without a reference genome has been carried out in switchgrass by defining linkage groups with the modulated modularity clustering (MMC) method (Stone and Ayroles 2009) referring to the genome of foxtail millet (Setaria italica) (F. Lu et al. 2013). A genetic map of wheat was

constructed by using the bin-mapping procedure with homozygous genotypes of a double-haploid population (Poland, Brown, et al. 2012). Each program for calling variants utilizes different models or algorithms to identify potential polymorphisms, therefore, multiple software programs need to be evaluated in order to identify the best

SNP caller for polyploids (Clevenger et al. 2015).

Linkage maps are important tools for map-based cloning, marker-assisted breeding, QTL identification, genome organization, and comparative genomics of important species. A number of linkage maps have been constructed for several grasses including pearl millet (Punnuri et al. 2016). However, so far, napiergrass SSR markers, genetic linkage map, or reference genome assembly are lacking. The purpose of this study was to survey the napiergrass genome and to construct a high-density genetic linkage map. Here, for the first time, we have surveyed whole genome sequences in napiergrass, developed SSR markers, and constructed high-density genetic map of napiergrass to investigate its genomic and genetic architecture.

Methods

Napiergrass Genome Survey

The genomic DNA of napiergrass cultivars Merkeron and UF1 was sequenced using Illumina Genome Analyzer and 454 GS-FLX. For Illumina reads, reads that contained more than 50% low-quality bases (Q20) were removed and adaptor sequences were trimmed. Quality and adapter trimming of 454 reads was done using default Newbler v2.8 (454 Life Sciences, Roche, Branford, CT) settings. Illumina reads were assembled using ABySS/1.3.4 (Simpson et al. 2009) with kmer size ranging from

25 to 60 at intervals of 5. The 454 reads were assembled using Newbler v2.8 (454 Life

Sciences, Roche, Branford, CT) with default parameters. The assemblies were

completed using CAP3 (Xiaoqiu Huang and Madan 1999). The largest 10 contigs of the assembly were selected for further analysis. Repeats on these contigs were masked using a comprehensive public repeat database compiled from TIGR plant repeats

(http://plantrepeats.plantbiology.msu.edu/), Plant miniature inverted-repeat transposable elements (P-MITE) database (http://pmite.hzau.edu.cn/django/mite/), MIPS Repeat

Element Database (http://mips.helmholtz-muenchen.de/plant/recat/), and Repbase from

RepeatMasker software (http://www.repeatmasker.org/). Unique repeats were extracted from this database by removing redundant repeats with 98% identity using CD-HIT/4.6

(Garrison and Marth 2012). SNPs and indels were called using FreeBayes/0.9.15

(Garrison and Marth 2012) excluding alleles with depth less than 20. The annotation of

SNPs was performed using SnpEff/4.0 (http://snpeff.sourceforge.net/) (Cingolani et al.

2012). In order to identify sequence similarity among the two genomes, clean reads from Illumina and 454 were aligned to the pearl millet genome v1 (Varshney et al. 2017) using bowtie2/2.2.5.

SSR Identification and Marker Development

The napiergrass assembly was used to identify SSR markers that contained repeat motifs ranging in length from 1 to 6 bp. The minimum number of repeats was 10 for Mono-, 6 for Di-and 5 for Tri-, Tetra-, Penta- and Hexa-. SSRs were analyzed based on their types, number of repeats, and percentage frequency of occurrences of each SSR motif. SSRs in napiergrass were detected using

MIcroSAtellite identification tool (MISA) (Thiel et al. 2003) and primers were developed using primer3 software (Rozen and Skaletsky 2000). SSR search results were input into scripts p3_in.pl and p3_out.pl in order to identify SSR primer pairs for napiergrass.

Plant Materials and DNA Extraction

A mapping population of 185 F1 hybrid progenies were developed from a cross between two napiergrass accessions (N122 and N190) described previously (Sinche

2013; Sinche et al. 2018). The 185 F1 hybrids were planted in the field at the Plant

Science Research and Education Unit (PSREU), Citra, Florida, along with the parental accessions.

Young and healthy leaf tissues were harvested from each individual of the mapping population. DNA extraction was done following the protocol described previously (Dellaporta, Wood, and Hicks 1983). The extracted DNA samples were run on a 2% agarose gel to check the quality and quantified with PicoGreen to meet the requirements of GBS. 185 F1 plants that were confirmed to be true hybrids using SSR markers (Sinche et al. 2018) were selected for GBS.

Genotyping-by-sequencing

GBS data was generated at the Institute of Biotechnology, Cornell University as described previously (Elshire et al. 2011). Briefly, DNA samples were digested with the restriction enzyme PstI followed by ligation of adapters, that consisted of Illumina sequencing primers and barcode adapters, to the DNA fragment ends. After ligation, 95 samples were combined into a pool and PCR amplification was performed to create a

GBS library and sequenced on Illumina HiSeq 2000.

Comparative Genomics

Unique tags of napiergrass from TASSEL de-novo UNEAK/3.0 (Glaubitz et al.

2014) pipeline were used for comparative genomic analysis. CD-HIT/4.6.4 (Weizhong Li and Godzik 2006) was used to cluster the tags. Genomes of rice (Osativa_323_v7.0),

Brachypodium (Bdistachyon_314_v3.0), maize (Zmays_284_AGPv3), sorghum

(Sbicolor_313_v3.0), foxtail millet (Sitalica_312_v2), switchgrass (Pvirgatum_273_v1.0), wheat (Taestivum_296_v2), Arabidopsis (Athaliana_167_TAIR9) were downloaded from

Phytozome v11 (https://phytozome.jgi.doe.gov/pz/portal.html). The barley genome

(ASM32608v1) was downloaded from Ensembl (http://www. ensembl.org). We used

BLASTN (BLAST v2.5.0) with the default settings and an e-value cutoff of 1×10-8 to blast the uniquely clustered tags of napiergrass against different genomes in order to find the percentage similarity of napiergrass reads among the various grass species.

Tags of 64 bp with > 80% identity and alignment length > 50 bp to the respective genomes were counted as a hit.

Sequence Analysis and SNP Calling

Raw data processing and SNP identification was performed using both de novo and reference-based approaches. Common software capable of calling SNPs de novo used in this research were TASSEL/3.0 UNEAK (Glaubitz et al. 2014), Stacks/1.24

(Catchen et al. 2013), and GBS-SNP-CROP 1.1 (Melo, Bartaula, and Hale 2016). For the reference based approach, pearl millet reference genome v1 (Varshney et al. 2017) was used. The reference genome consists of seven pseudomolecules. Six different reference based pipelines were evaluated to call SNPs viz., TASSEL 4.3 (Glaubitz et al.

2014), Stacks 1.24 (Catchen et al. 2011), GBS-SNP-CROP 1.1 (Melo, Bartaula, and

Hale 2016), SAMtools 1.2 mpileup (H. Li et al. 2009), FreeBayes 0.9.21 (Garrison and

Marth 2012), and GATK 3.3 (McKenna et al. 2010). Parameters used in each software are provided in Table 2-1. Sequence variants called were filtered with a minimum depth of 48 per sample.

Linkage Map Construction

QC-filtered SNPs were further filtered by the following standards for map construction: 1) markers must be genotyped in at least 180 individuals; 2) Individuals with over 10% missing data were discarded; and 3) Redundant markers were removed by standard of similarity = 1. For each parental map construction, only single dose markers were used. Markers segregating at a distorted Mendelian ratio (expected ratio for ‘lmxll’ type and ‘nnxnp’ type is 1: 1, χ2 test, 0.001 < P < 0.05) were marked. The single dose markers from the maternal and paternal parent were analyzed separately using JoinMap 4.1 (van Ooijen 2006) and outcross pollinated family (CP) was selected as the population type. Markers that were heterozygous in N122 and homozygous in

N190 (‘lmxll’ type) were selected to build N122 linkage groups. Markers that were heterozygous in N190 and homozygous in N122 (‘nnxnp’ type) were selected to build

N190 linkage groups. The linkage groups were built using regression mapping algorithm, with a minimum logarithm of odds (LOD) value at 20, and a maximum recombination frequency at 0.40. Marker positioning calculation was performed with a goodness-of-fit jump at 5, followed by a “ripple” procedure (value =1). Map distances were estimated using the Kosambi mapping function. Genetic distance between SDR markers were corrected using DistortedMap (S. Q. Xie, Feng, and Zhang 2014).

Linkage maps were drawn with MapChart (Voorrips 2002). For integrated map construction, markers that were heterozygous in both parents (‘hkxhk’ type) were selected to build combined linkage groups. Markers segregating at distorted Mendelian ratio (expected ratio for ‘hkxhk’ type is 1:2:1, χ2 test, 0.001< P < 0.05) were marked. The retained markers were then added with the markers from male and female parents to construct a combined map. The linkage groups were built using regression mapping 45

algorithm, with a minimum logarithm of odds (LOD) value at 20, and a maximum recombination frequency of 0.40. Other parameters were the same with linkage map construction above. Regions showing segregation distortion (0.001< P <0.05) with more than three adjacent loci were marked as SDR regions (Paillard et al. 2003; Z. Zhang et al. 2016).

Comparison Between Napiergrass and Pearl Millet Genome

Consensus sequence of mapped markers from TASSEL de-novo UNEAK were used to compare with the reference genome of pearl millet with same parameters

(BLASTN defaults with an e-value cutoff of 1x10-8). Markers that showed significant hits to the genome sequence and /or gene models of pearl millet with > 80% identity and alignment length > 50 bp were extracted and used for comparative genomics study. A circos plot was drawn using circos/0.69-2 (Krzywinski et al. 2009).

Results

Napiergrass Genome Survey

Approximately 211 million raw reads from Illumina and 97 thousand raw reads from 454 sequencing were subjected to a sequence quality check. After filtration and trimming, 161 million clean Illumina reads and 96,000 clean 454 reads were aligned to the pearl millet genome v1 (Varshney et al. 2017). A total of 62.5 million (38.8%) reads were able to align with the pearl millet genome. Polymorphisms were detected between the napiergrass and pearl millet aligned reads, of which 619,708 SNPs and 24,135 indels were identified. Most of the sequence variations (58.7% SNP) were in intergenic regions (Figure 2-1). The clean reads were assembled into 113,789 contigs with a total size of 44.5 Mbp and a N50 of 435 bp and a GC content 43.45%. The largest 10 contigs of the sequence assembly, which ranged from 8,506 to 25,329 bp in length, were

selected as representative napiergrass genome fragments. The repeat content of the 10 longest contigs ranged from 5% to 90% with an average of 27.5% and a total of 164 repetitive elements (Table 2-2). Two contigs had no hits in the pearl millet genome due to a high repeat content (over 80%). The rest of the contigs had one or more large hits

(>500 bp) to the pearl millet genome. The micro-synteny showed mostly collinear relationship between the genomes of the two species (Figure 2-2). However, chromosome inverted duplications were also observed in the pearl millet genome

(Figure 2-3), indicating that the chromosome rearrangement occurred after the speciation of napiergrass and pearl millet. The length of stringently (>500 bp and >80% sequence similarity) aligned regions accounted for 36.3% of the examined contig sequences of napiergrass (Table 2-3). The total length of the alignment was 25.1% higher in pearl millet than in napiergrass aligned regions.

From the assembled napiergrass survey sequences, 5,339 SSRs were identified.

Mono- type repeats were most common in napiergrass, followed by Tri-, Di- and Tetra- type repeats (Table 2-4). From these identified SSRs, 1,926 were successfully used for primer design (Table 2-5). All of the primer sequences aligned to the assembly of napiergrass and 89% of the primers were uniquely aligned. On the other hand, the overall alignment rate of the primers with pearl millet genome v1 (Varshney et al. 2017) was 31% with 15% uniquely aligned. These SSR primers will undoubtedly serve as an abundant resource for molecular markers in napiergrass.

Genotyping-by-sequencing

To construct the linkage map for napiergrass, an F1 bi-parental population was developed, which consisted of 185 true hybrid individuals (Sinche et al. 2018). These hybrids were subjected to GBS. A total of 549 million raw reads were generated. After 47

trimming and filtering, 512 million high quality reads were retained. The average number of reads per sample was 2.6 million and ranged from 44 thousand to 5.4 million reads per sample. In silico digestion of the pearl millet genome v1 (Varshney et al. 2017) with

PstI yielded DNA fragments in the range of 170-350 bp, which suggest that an estimated average depth for the mapping population was 36.5 reads per locus per sample (Figure 2-4), which should allow us to call the SNPs confidently at most of the loci.

A total of 695,602 unique tags were identified from the clean reads generated from the mapping population by using the TASSEL de-novo UNEAK pipeline. These tags were further clustered into 182,934 non-redundant tags by CD-HIT. To examine the sequence similarity between napiergrass and other grass species, we aligned the non-redundant tags of napiergrass against several grass species with complete genome sequences including rice (Oryza sativa) (Osativa_323_v7.0), Brachypodium

(Bdistachyon_314_v3.0), maize (Zmays_284_AGPv3), sorghum (Sbicolor_313_v3.0), foxtail millet (Sitalica_312_v2), switchgrass (Pvirgatum_273_v1.0), wheat

(Taestivum_296_v2), pearl millet v1(Varshney et al. 2017) , and barley (ASM32608v1), with Arabidopsis (Athaliana_167_TAIR9) as an outgroup control. The results showed that the percentage of napiergrass sequence tags aligned to these grass species ranged from 2.6% to 37.9% for barley and pearl millet genome, respectively (Table 2-6), indicating a relatively close relationship between napiergrass and pearl millet.

SNP Calling by Various SNP Callers

Three de-novo SNP calling pipelines, TASSEL-UNEAK, Stacks, and GBS-SNP-

CROP identified 10,799, 6,871, and 4,521 SNPs, respectively. Reference based pipelines were also applied by using pearl millet v1 (Varshney et al. 2017) as the reference genome. However, the alignment rate was relatively low due to the differences between the napiergrass and pearl millet genomes. The percentage of clean reads aligned to the pearl millet genome using Bowtie2 ranged from 5.60% to 44.62% with an average of 39.68%. Two samples had a small number of sequences (< 10% of the average number of sequences per sample) and also the lowest percentage of uniquely mapped reads (Table 2-7, Figure 2-5). Therefore, these samples were removed from linkage map construction. Six different reference-based pipelines were employed to call SNPs viz., TASSEL 4.3 (Glaubitz et al. 2014), Stacks 1.24 (Catchen et al. 2011), GBS-SNP-CROP (Melo, Bartaula, and Hale 2016), SAMtools 1.2 mpileup (H.

Li et al. 2009), FreeBayes 0.9.21 (Garrison and Marth 2012), and GATK 3.3 (McKenna et al. 2010). TASSEL 4.3, Stacks, and SAMtools identified 7,326, 4,920, 27,082 SNPs, respectively in the mapping population, whereas FreeBayes, GBS-SNP-CROP, and

GATK that can handle ploidy identified 25,193, 2,906 and 197,475 SNPs, respectively.

The six reference-based SNP callers concordantly called only 11 SNPs (Figure 2-6, only five programs are shown in figure due to Venn-diagram display limitations) and

207,391 non-redundant SNPs.

Genetic Linkage Map Construction

From a total of 549,944 SNPs called by both reference based and de-novo pipelines, 287,093 SNPs were filtered for further analysis. Out of these, a total of 18,286 single-dose SNPs were genotyped in more than 180 progenies. Three individuals with

more than 10% missing sites were removed from further analysis. For linkage map construction of each parental line, only the SNPs showing heterozygous in one parent but homozygous in the other parent were selected. A total of 3,276 loci were heterozygous in female parent but homozygous in male parent and segregated with an expected ratio of 1:1 in the population, thus can be used for female parent linkage map construction. Similarly, 3,417 loci were heterozygous in male parent but homozygous in female parent and segregated with an expected ratio of 1:1 in the population, thus can be used for male parent linkage map construction. For the female parental line, a total of 1,606 SNPs were grouped and 899 loci were mapped on 14 linkage groups with a total length of 1,555.17 cM averaging 1 marker every 1.72 cM (Figure 2-7). Inclusion of segregation distorted (SD) markers increased the genetic distance of the female map by

28.13%. For the male parent, a total of 1,509 markers were grouped into 14 linkage groups and 1,073 markers were mapped onto these 14 linkage groups with a total length of 1,939.19 cM averaging 1 marker every 1.80 cM (Figure 2-8). Inclusion of SD markers increased the total genetic distance of the male map by 38.41%.

A combined linkage map containing markers that segregated from both female and male parents was constructed, which can facilitate future QTL mapping of the population. To construct a combined linkage map, the markers showing heterozygous on both parents in addition to male-parent heterozygous and female-parent heterozygous markers were used. Therefore, a parent-averaged combined map was constructed by using 378 heterozygous markers for both parents that segregated in a

1:2:1 ratio in the population, in combination with 3,417 male-parent heterozygous and

3,276 female-parent heterozygous markers. In total, 4,058 markers were grouped into

14 linkage groups out of which 1,913 markers were mapped. The final composite linkage map spanned a length of 1,410.10 cM with an average of 0.73 cM between markers. The largest linkage group was Linkage group 2 (LG 2), which spanned 142.40 cM and contained 170 markers (Table 2-8). Length of each linkage group ranged from

70.18 cM to 142.40 cM and density ranged from 0.88 to 1.77 markers per cM (Figure 2-

9, Table 2-8, Figure 2-10). Results of the χ2 test indicated that 114 (6.06%) of the 1,879 markers showed significant segregation distortion (0.001

Among the different reference-based SNP callers, GATK called the highest number of SNPs (197,475) followed by SAMtools and FreeBayes (Table 2-9). After accounting for segregation ratio and missing data, SAMtools retained the largest number of SNPs followed by TASSEL de-novo UNEAK. However, when considering the total number of markers mapped on the combined linkage groups, TASSEL de-novo

UNEAK showed the highest percentage of SNPs mapped followed by Stacks (Table 2-

9).

Comparison Between Genomes of Napiergrass and Pearl Millet

Sequence tags of the markers that mapped on napiergrass linkage groups were extracted and compared to the pearl millet genome. Among the 1,156 TASSEL de-novo

UNEAK tags positioned on the combined map, 663 were found to have significant sequence similarities to the genome sequence of pearl millet. Considerable collinearity was observed between the napiergrass and pearl millet genomes (Figure 2-11). For each pearl millet pseudomolecule, two corresponding regions in the linkage groups

(LGs) of napiergrass genome were identified (Figure 2-11, Figure 2-12). However, some 51

pearl millet genomic regions had more than two corresponding regions on napiergrass genome. For example, pseudomolecule 3 of pearl millet had regions corresponding to three linkage groups LG03, LG12, and LG14 of napiergrass indicating possible chromosomal rearrangement between the two species after speciation (Figure 2-11,

Figure 2-12).

Discussion

Despite its importance as a forage grass and its enormous potential as a biofuel crop, molecular, genetic, and genomic studies have been severely limited in napiergrass. Currently, there was no equivalent genome sequence in the public domain to be used as a reference for napiergrass. In this study, an initial comparison between the napiergrass survey sequences to 10 available grass genomes revealed that napiergrass genomic sequences had the highest similarity with the pearl millet genome, which could be explained by the presence of the A’A’ genome of napiergrass that is homologous to the AA genome of pearl millet. Consequently, in this study we have utilized pearl millet genome v1 (Varshney et al. 2017) as a reference for SNP calling and also performed de-novo SNP calling without a reference genome. A total of 38.8% of the napiergrass reads aligned to the pearl millet genome using Bowtie 2, which performed better over BWA, another popular aligner (Langmead and Salzberg 2012;

Yang, Song, et al. 2017). The large portion of unaligned reads might be from the B genome or the divergent chromosome regions of A genome between the two species.

From the genome survey comparison, the total length of all the alignments of napiergrass reads was 25.1% longer in pearl millet indicating genic duplication or expansion in pearl millet and genomic rearrangements between the two species during evolution from their ancestral genome. This is consistent with a previously reported 52

genomic in situ hybridization, which verified that the pearl millet genome A was 24% larger compared to the chromosomes of genome A’ of napiergrass (Reis et al. 2014).

For the 10 longest contigs in our assembly, average repeat content (27%) was lower than reported from other grasses including sorghum (61%) (Paterson et al. 2009), maize (85%) (Schnable et al. 2009), foxtail millet (46%) (G. Zhang et al. 2012), rice

(43.3%) (J. Yu et al. 2002), and pearl millet (77%) (Varshney et al. 2017). Low repeat content in napiergrass could be attributed to the loss of genomic sequences after hybridization. Rearrangements and loss of genomic sequences are common events after hybridization (Kellis, Birren, and Lander 2004). Similar to other plant genomes, long-terminal repeat (LTR) retrotransposons comprised the most abundant class (62.19

%) of repeats (Table 2-2). Significant relationships between napiergrass, pearl millet, and P. squamulatum suggested their common origin and it was inferred that napiergrass and pearl millet had concomitantly diverged from a common ancestor (Reis et al. 2014; Martel et al. 2004; Martel et al. 1997) and the origin of napiergrass occurred at the interspecific hybridization event, by combining genome A of the ancestor with genome B of a still unknown second ancestor (Reis et al. 2014). Our study showed that the napiergrass genome had considerable microcolinearity with the pearl millet genome showing evidence of their relatedness and shared ancestry. Chromosome inverted duplications on pseudomolecule 3 of pearl millet showed possible rearrangement after speciation of napiergrass and pearl millet. Two corresponding regions on the napiergrass linkage groups for each pearl millet chromosome corroborate the hypothesis that these two genomes evolved from a common ancestor.

We developed a limited genomic assembly of napiergrass based on Illumina and

454 sequences. Nearly two thousand SSR markers were developed, which could be immediately useful for applications in napiergrass breeding and genetics. With the advancement of NGS, high throughput NGS-enabled genotyping technologies are becoming readily accessible. Yet, SSR markers remain as a popular tool for genetic studies, variety identification, monitoring of seed purity, and hybrid quality. They are particularly important in laboratories which have limited resources and lack access to

NGS facilities or bioinformatic expertise. To our knowledge, this is the first study in napiergrass where SSR markers were developed based on napiergrass genome survey.

A genetic linkage map is an important tool to reveal the genome structure and to identify marker-trait associations (Cai et al. 2015) which ultimately help in MAS (F. Lu et al. 2013) to improve precision of selection. In this study, we used the GBS approach to construct a combined high-density linkage map that spanned 1,473.9 cM with 1,917 markers on 14 linkage groups, which is a very critical tool for further genetic and genomics studies of napiergrass. GBS has been extensively used for genotyping many diploid organisms, however, SNP calling from the NGS data in allotetraploids like napiergrass is particularly challenging due to existence of highly similar homeologous copies, one corresponding to A genome and the other to B genome (Nagy et al. 2013).

Therefore, different strategies have been devised to construct linkage map in allopolyploids. For example only uniquely aligned reads (single copy) were considered for SNP calling and subsequent map construction (Trick et al. 2009; X. Zhou et al. 2014) to avoid the collapsed alignment of homoeologous reads due to low divergence, recent

polyploidization event, and severe domestic bottlenecks (Pandey et al. 2012). SNP calling in allotetraploid Brassica napus L. (rapeseed; 2n = 4x =38; AACC) was done by utilizing only uniquely mapped reads (single copy) and a read depth minimum of three to four reads at each potential SNP (Trick et al. 2009). Linkage map construction in zoysiagrass (Zoysia matrella) was performed by utilizing single-dose markers after calling SNPs using the maximum likelihood method in Stacks (Xiaoen Huang et al.

2016). Similarly, single dose markers from TASSEL de-novo UNEAK were used to construct linkage maps in prairie cordgrass (Spartina pectinate)(Crawford et al. 2016).

In this study, we applied multiple SNP callers and strategies to maximize SNP calling for linkage map construction for napiergrass. In the final combined genetic map, the number of markers identified by different software varied dramatically. GATK called the highest number of SNPs followed by SAMtools and FreeBayes initially. Both GATK and SAMtools apply Bayesian method to compute the posterior probability for each possible genotype and then choose the genotype with the highest probability as the consensus genotype (X. Yu and Sun 2013). GBS-SNP-CROP and TASSEL showed a low matching percentage, which is similar to results from previous research (Melo,

Bartaula, and Hale 2016). The number of useful markers for linkage group construction was the highest in SAMtools (47.75%) followed by TASSEL de-novo UNEAK (35.68%).

However, the TASSEL de-novo UNEAK pipeline had the highest number of markers mapped on the linkage groups (60.43%) followed by Stacks (13.43%). This indicated that the network-based SNP discovery in TASSEL de-novo UNEAK and UStacks pipeline (Kim et al. 2015) could be efficiently utilized for constructing linkage maps in non-model species. Even though TASSEL was primarily designed for diploids, it is

powerful enough to give a large number of mapped markers compared to other programs that handle polyploidy like FreeBayes, GATK, or GBS-SNP-CROP.

The SNP markers were relatively evenly distributed among the linkage groups with more than 97.45% of marker interval being less than 5 cM. To our knowledge, this linkage map with an average inter-marker distance of 0.7 cM is the first genetic linkage map of napiergrass to date. A study based on an interspecific population of a cross between pearl millet and napiergrass has been previously reported to link RAPD markers with biomass related traits in Pennisetum (Smith et al. 1993). The large number of markers and their even distribution in our study facilitate full-scale map coverage.

Few regions where the interval space was > 5 cM might be due to stretches of large repeats or due to low coverage sequencing of GBS (Poland and Rife 2012; Mathew et al. 2014). Segregation distortion is regarded as a potential evolutionary force and including these markers for linkage map construction could increase genome coverage as well as benefit QTL mapping (S. Xu 2008; D. R. Taylor and Ingvarsson 2003).

Including SDR markers and correcting for bias led to an increase in genetic distance between distorted markers (S. Q. Xie, Feng, and Zhang 2014). The deviation from expected Mendelian ratio shows disturbances in the transmission of genetic information from one generation to the next and can be caused by chromosome loss or rearrangements, genetic load, gametic selection, zygotic selection, or both (Faris,

Laddomada, and Gill 1998; Karkkainen, Koski, and Savolainen 1996; Bodénès et al.

2016). Napiergrass generally outcross through wind pollination that could result in high levels of gene flow leading to genetic load. The assignment of napiergrass linkage

groups according to the pearl millet genome allows for future fine mapping and QTL analysis.

In summary, this study reports for the first time a high-density genetic linkage map using NGS-derived SNP markers, as well as the development of SSRs from napiergrass genomic sequences. The napiergrass genome showed considerable collinearity with the pearl millet genome and the genetic map contains 14 linkage groups with low inter-marker interval. The results will be useful for future molecular breeding programs such as identification of QTLs for important traits as well as MAS for the genetic improvement of napiergrass and comparative genomics. These resources will play a critical role in future whole genome sequencing projects and leveraging molecular breeding of napiergrass.

Table 2-1. Parameters used for SNP calling for each software. Reference based Parameters Remarks. [defaults] TASSEL 4.3 -c 5 Min. number of times a tag must be present to be output <5> [1] -mnMAF 0.01 Min. minor allele frequency <0.01>[0.01] -mnMAC 100000 Min. minor allele count <100000>[10] (SNPs that pass either -mnMAF or -mnMAC will be output) -misMat 2 Threshold genotypic mismatch rate above which the duplicate SNPs won’t be merged <2>[0.05] -callHets When two genotypes at a replicate SNP disagree for a taxon, call it a heterozygote Stacks -A CP CP type for genetic map -m 3 Min. number of identical, raw reads required to create a stack <3>[3] GBS-SNP-CROP -l 30 Trimmomatic LEADING parameter -sl 4:30 Trimmomatic SLIDINGWINDOW parameter -tr 30 Trimmomatic TRAILING parameter -m 32 Trimmomatic MINLEN parameter -rl 100 Raw GBS read length -pl 32 Min. length required after merging to retain read -p 0.01 p-value for PEAR -id 0.93 Nucleotide identity value required for USEARCH read clustering -Q 30 Phred score base call quality -q 0 Alignment quality -f 0 SAMtools flags -F 2308 SAMtools flags -mnHoDepth0 11 Min. depth required for calling a homozygote when the alternative allele depth = 0 -mnHoDepth1 48 Min. depth required for each allele when calling a heterozygote -mnHetDepth 3 Min. depth required for each allele when calling a heterozygote

Table 2-1. Continued Reference based Parameters Remarks. [defaults] -altStrength 0.9 Across the pop. For a given putative bi-allelic SNP, this alternate allele strength is the minimum proportion of non-primary allele reads that are the secondary allele -mnAlleleRatio 0.1 Min. required ratio of less frequent allele depth to more frequent allele depth -mnCal 0.75 Min. acceptable proportion of genotyped individuals to retain a SNP -mnAvgDepth 4 Min. avg. depth of an acceptable SNP -mxAvgDepth 200 Max avg. depth of an acceptable SNP SAMtools mpileup -uf Default FreeBayes -C 2 --min-alternate-count Require at least <2> observations supporting an alternate allele within a single individual in order to evaluate the position [1] -p 4 --ploidy <4> [2] --use-best-n- Evaluate only the best N SNP alleles ranked by alleles 4 sum of supporting quality scores [all]

--min-coverage 5 Require at least <5> coverage to process a site [0] GATK -T Call SNPs and indels on a per-locus basis UnifiedGenotyper -stand_call_conf The min. phred-scaled confidence thresholds at 30 which variants should be called <30> [30] -stand_emit_conf The minimum phred-scale confidence threshold 10 at which variants should be emitted (and filtered with LowQual if less than the calling threshold) <10> [30] -ploidy 4 Ploidy <4> [2] -mbq 20 Minimum base quality required to consider a base for calling -glm BOTH Genotype likelihoods calculation model includes SNPs and INDELs [SNP]

Table 2-1. Continued de-novo based Parameters Remarks. Defaults [] UNEAK -e PstI Restriction enzyme used -c 5 Min. count of a tag must be present to be output [5] -e 0.03 Error tolerance rate in the network filter [0.03] -mnMAF 0.05 Min. minor allele frequency [0.05] -mxMAF 0.5 Max. minor allele frequency [0.5] -mnC 0 Min. call rate (proportion that how many taxa are covered by at least one tag) -mxC 1 Max. call rate [1] Stacks -m 3 Min. number of identical, raw reads to create a stack -M 2 No. of mismatches allowed between loci when processing a single individual [2] -n 1 No. of mismatches allowed between loci when building the catalog [1] -t Remove, or break up, highly repetitive RAD-Tags in the ustacks program GBS-SNP-CROP Same as reference based except, script 3 settings not required

Table 2-2. Repetitive elements present in the napiergrass genome. Transposable Element Count DNA transposon

Tc1/Mariner 8 hAT 20 PIF/Harbinger 16

EnSpm 1 CACTA 4 Polinton 1

LTR Retrotransposon LTR 37 LTR/Copia 5

LTR/Gypsy 5 Copia 4 Retrotransposon 3

Retroelement 1 Gypsy 2 Non-LTR Retrotransposons

L1 2 Pseudogene tRNA 1

rRNA 8 rDNA-like 3

Others Mutator 15 Telomeric 4 MobileElement 2 Micro-like sequence 1 Low-complexity 3 Simple Repeats 12 Unspecified 6 Total 164

Table 2-3. The sequence alignment of ten napiergrass sequence contigs to the pearl millet genome. Napiergrass Size of GC Repeat Corresponding Number Number of Alignment Alignment Sum of contig contig content content pearl millet of hits predicted length in length in expanded (bp) (%) (%) pseudomolecule above genes napiergrass pearl millet length in 500bp pearl millet No.

1 Contig1 8,506 55.3 80.3 Pg5 0 0 / / / 2 Contig3434 9,653 42.5 9.8 Pg3 5 0 5,407 5,384 -23 3 Contig5516 11,920 44.9 17.5 Pg3 11 1 6,776 13,509 6733 4 Contig5578 8,558 42.8 5.9 Pg2 3 0 5,948 6,487 539 5 Contig5588 8,595 44.6 14 C26927002 1 1 591 591 0 6 Contig5729 9,088 46 20.4 Pg5 3 0 2,250 2,246 -4 7 Contig5798 8,651 43.1 5.3 C27370090 1 0 1,385 1,385 0 8 Contig5878 25,329 43.4 12.8 Pg6 4 0 5,692 5,676 -16 9 Contig5902 13,801 46.1 16.4 Pg6 8 0 6,209 7,596 1387 10 Contig6139 14,330 42.2 93.1 Pg7 0 0 / / / Total 10 118,431 45.09 27.5 8 36 2 34,258 42,874 8,616

Table 2-4. Frequency of classified repeat types (considering sequence complementary) in napiergrass. Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total A/T - - - - - 1045 338 173 76 49 34 17 9 10 3 12 1766 C/G - - - - - 116 71 36 26 11 7 7 6 2 1 283 AC/GT - 118 43 27 18 14 6 2 1 1 1 2 1 9 243 AG/CT - 250 137 82 45 31 19 14 13 9 6 3 5 3 2 6 625 AT/AT - 180 81 55 40 17 4 5 4 8 5 1 4 12 416 CG/CG - 30 10 1 41 AAC/GTT 45 15 14 6 1 2 1 84 AAG/CTT 260 85 46 27 13 13 8 7 12 3 2 476 AAT/ATT 62 31 6 4 5 2 1 2 1 1 2 117 ACC/GGT 86 35 15 7 3 1 2 1 1 151 ACG/CGT 45 3 7 1 56 ACT/AGT 35 12 4 2 1 54 AGC/CTG 131 53 30 15 4 3 236 AGG/CCT 129 34 10 5 1 2 1 182 ATC/ATG 86 38 13 13 4 3 3 3 1 1 1 166 CCG/CGG 157 39 18 3 1 1 219 AAAC/GTTT 2 2 AAAG/CTTT 14 5 3 2 2 26 AAAT/ATTT 19 1 2 1 1 2 26 AACC/GGTT 3 1 4 AAGC/CTTG 1 1 AAGG/CCTT 1 1 AATC/ATTG 1 1 AATG/ATTC 2 1 1 4 AATT/AATT 1 1 ACAG/CTGT 5 1 1 7 ACAT/ATGT 6 2 1 1 1 5 1 1 4 22 ACCC/GGGT 1 1 ACCG/CGGT 1 1

Table 2-4. Continued Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total ACCT/AGGT 1 1 ACGC/CGTG 2 2 ACGG/CCGT 1 1 ACTC/AGTG 2 2 ACTG/AGTC 1 1 AGAT/ATCT 7 2 2 1 1 13 AGCC/CTGG 1 1 2 AGCG/CGCT 1 1 AGCT/AGCT 1 1 AGGC/CCTG 1 1 2 AGGG/CCCT 2 2 ATCC/ATGG 6 1 1 1 9 ATGC/ATGC 3 3 CCGG/CCGG 2 2 AAAAC/GTTTT 2 2 AAAAG/CTTTT 1 1 2 AAAAT/ATTTT 1 1 AAACC/GGTTT 1 1 AAAGG/CCTTT 1 1 2 AAAGT/ACTTT 1 1 AAATC/ATTTG 1 1 AACAG/CTGTT 1 1 AACAT/ATGTT 1 1 AACCT/AGGTT 1 1 AACTG/AGTTC 1 1 AAGAG/CTCTT 1 1 AAGAT/ATCTT 1 1 AAGCT/AGCTT 1 1 AAGGG/CCCTT 1 1

Table 2-4. Continued Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total AATAC/ATTGT 1 1 AATAG/ATTCT 2 1 1 4 AATCC/ATTGG 1 1 AATCT/AGATT 1 1 2 AATGG/ATTCC 1 1 AATGT/ACATT 1 1 ACACC/GGTGT 1 1 2 ACACG/CGTGT 2 2 ACCAG/CTGGT 1 1 ACCGC/CGGTG 1 1 ACGGC/CCGTG 1 1 ACGTC/ACGTG 1 1 ACTAT/AGTAT 1 1 2 ACTCC/AGTGG 2 2 AGAGG/CCTCT 2 2 AGATG/ATCTC 1 1 2 AGCCG/CGGCT 2 2 AGCCT/AGGCT 1 1 AGGCC/CCTGG 1 1 ATCCC/ATGGG 1 1 AAAAAG/CTTTTT 1 1 AAATCG/ATTTCG 1 1 AACCCG/CGGGTT 1 1 AACTCC/AGTTGG 1 1 AAGAGC/CTCTTG 1 1 AAGAGG/CCTCTT 2 2 AAGATG/ATCTTC 2 1 2 5 AAGCAC/CTTGTG 1 1 AAGGAG/CCTTCT 3 3

Table 2-4. Continued Repeats 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ≥20 Total AAGTAG/ACTTCT 1 1 AATAGG/ATTCCT 1 1 AATAGT/ACTATT 1 1 AATCCC/ATTGGG 1 1 AATCGG/ATTCCG 1 1 ACCAGC/CTGGTG 1 1 ACCATC/ATGGTG 1 1 2 ACCCCG/CGGGGT 1 1 ACCGTG/ACGGTC 1 1 ACGGGG/CCCCGT 1 1 ACTGCT/AGCAGT 1 1 2 AGATAT/ATATCT 1 1 AGATGG/ATCTCC 1 1 AGCAGG/CCTGCT 1 1 AGCATC/ATGCTG 1 1 AGCATG/ATGCTC 1 1 AGCCGG/CCGGCT 1 1 AGCCTG/AGGCTC 1 1 Total 1172 953 451 252 142 1263 459 243 136 84 56 31 26 16 6 49 5339

Table 2-5. Primer pairs developed for napiergrass SSR markers. Available at https://www.nature.com/articles/s41598-018-32674-x#Sec18

Table 2-6. Summary of the alignment of non-redundant tags of napiergrass (Cenchrus purpureus) to the available genomes of different species. Genome used (species name) Number of tags Percentage of tags with blast hits with blast hits (%)

Arabidopsis (Arabidopsis thaliana) 120 0.07

Purple false brome (Brachypodium 6,029 3.30 distachyon)

Barley (Hordeum vulgare) 4,751 2.60

Rice (Oryza sativa) 6,879 3.76

Pearl millet (Pennisetum glaucum) 69,385 37.93

Switchgrass (Panicum virgatum) 24,163 13.21

Sorghum (Sorghum bicolor) 14,654 8.01

Foxtail millet (Setaria italica) 40,849 22.33

Wheat (Triticum aestivum) 6,459 3.53

Maize (Zea mays) 11,972 6.54

Table 2-7. Alignment of individual napiergrass reads using Bowtie2. Number of raw reads Clean reads (retained % of reads Total reads mapped Uniquely mapped % uniquely by Stacks) retained to pearl millet mapped

Min. 44,858 20,350 45.37% 16,658 1,140 6.00% Max. 5,046,114 4,739,605 97.53% 3,022,185 1,902,058 45.00% Avg. 2,893,836.07 2,696,528.11 92.71% 1,727,194.04 1,077,802.96 39.68% Total 549,828,854 512,340,341 93.18% 328,166,867 204,782,562 39.97%

Table 2-8. Summary of the combined linkage map of napiergrass and the percentage of gaps less than 5 cM in male and female parent linkage maps. Napier grass Pearl millet Number of Mapped Unmapped Length Density Combined Female Male linkage group syntenic grouped markers markers (cM) (markers map map Gaps map pseudomolecule markers per cM) Gaps <=5 <= 5cM (%) Gaps cM (%) <=5 cM (%) 98.15 93.98 93.75 LG01 PM01 411 163 248 109.33 1.49 97.63 89.19 92.45 LG02 PM06 378 170 208 142.40 1.19 98.06 96.97 85.71 LG03 PM03 360 156 204 89.81 1.74 99.21 95.51 85.53 LG04 PM05 339 127 212 74.83 1.70 98.34 96.55 96.36 LG05 PM02 324 182 142 105.50 1.73 96.09 96.23 96.83 LG06 PM04 300 129 171 112.53 1.15 96.94 89.8 84.72 LG07 PM07 279 99 180 96.84 1.02 97.48 84 89.86 LG08 PM06 279 120 159 98.45 1.22 92.63 94.64 95.24 LG09 PM02 278 96 182 108.70 0.88 98.89 92.65 75.76 LG10 PM01 254 181 73 102.49 1.77 98.57 91.53 88.06 LG11 PM07 237 141 96 97.35 1.45 97.24 89.29 90.11 LG12 PM03 230 146 84 105.72 1.38 98.00 83.67 94.12 LG13 PM05 224 101 123 70.18 1.44 97.03 89.09 90.91 LG14 PM03 165 102 63 96.00 1.06 (97.45) (91.65) (89.96) Total (average) 4,058 1,913 2,145 1,410.10 (1.37)

Table 2-9. Summary of napiergrass single nucleotide polymorphism (SNP) markers mapped on the combined linkage map using 9 different software pipelines. Software Number of Total SNPs No. of SNPs on Percentage of SNPs called used for map map SNPs on the construction map (%)

FreeBayes 25,193 6 0 0.00

GATK 197,475 52 5 0.26

SAMtools 27,082 3,377 151 7.89

GBS-SNP-CROP 2,906 115 52 2.72

TASSEL 7,326 116 56 2.93

Stacks 4,920 447 257 13.43

GBS-SNP-CROP 4,521 96 51 2.67 de-novo

Stacks de-novo 6,871 339 185 9.67

TASSEL de-novo 10,799 2,523 1,156 60.43 UNEAK

Total 287,093 7,071 1,913

70% Intergenic 60% Up 50% 5' UTR Exon 40% Donor 30% Intron Acceptor 20% Exon 10% 3' UTR 0% Down Intergenic

Figure 2-1. Sequence variation for SNPs called in various regions of the pearl millet genome.

Figure 2-2. Micro-collinearity between contigs from napiergrass to the pearl millet genome.

Figure 2-3. Inversion duplication between napiergrass and pearl millet (shown in bottom figure).

Number Number of samples 10

0 5

60 10 15 20 25 30 35 40 45 50 55 65 70 75 80 85 90 95

100 More Estimated coverage (X)

Figure 2-4. Estimated coverage of PstI restriction sites in the pearl millet genome.

100 90 80 70 60 50 40

30 Number Number of samples 20 10 0 <6 // 34 36 38 40 42 44 46 48 Percentage of uniquely mapped reads (%)

Figure 2-5. Histogram of uniquely mapped reads to the pearl millet genome.

Figure 2-6. Venn diagram showing concordant napiergrass SNPs called by five reference-based SNP callers, SAMtools, GBS-SNP-CROP, GATK, FreeBayes, and TASSEL. Numbers in parenthesis after the program name shows the total number of SNPs called by each program.

Napiergrass: Female Parent Linkage Map 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 dT20387 S6_80432248 dT28623 S1_6210777 S2_16737601 S4_43915384 S7_182881700 dT50316 dT14190 dT29619 dT49930 dT25904 dT6944 S7_177490027 dT41575 dT35743 S2_16737644 S4_43915389 S4_131830044 G4_131830067 dT1702 dT3027 S4_131830067 dT44709 dT909 dT30402 dT37150 S4_131830066 dT43814 C1_10467974 dT39834 S4_131830063 dT6724 5 S5_128073888 dT12313 dT38795 dT53129 dT5178 dC_93403 dC_63791 dT42418 dT22876 dT8647 S3_64306230 C6_237238320 S2_174624948 dT20158 dT12269 dT7257 dT43720 dT3328 S1_6210761 dT12265 dT27092 C4_131830077 dT49252 dT28282 dC_61912 dC_55022 dT51744 dT21137 10 dT30456 dT42723 dT38918 dT21313 dT47966 dT22567 S3_200493878 dT24515 dT30583 dT16886 S2_24418753 dT16746 dT47917 dT26116 dT37862 dT27350 S7_112269 dT53891 dT12948 dT5844 dT38883 dT17058 dT17924 G7_212623362 dT10917 dT16764 dT46838 S2_24418742 dT3588 dT36340 dT10040 dT9730 TS5_114098156 dT46230 dT15762 15 dT35147 dT165 dT42842 S4_12461988 dT29992 dT26884 dT24050 dT19814 dT16773 dT910 dT49010 dT13195 S3_58228915 dT19311 dT15691 dT7436 dT10739 dT621 dT38555 dT40995 dT38665 dT26620 dT51352 dT2473 dT35970 dT17556 dT20060 dT17325 S7_89250999 dT38857 dT24731 dT15817 dT47411 S4_88888650 dT43326 dT12719 dT1951 dT7946 dT24472 dC_23137 dT17578 dT19291 dT3341 dT51987 dT14651 dT3578 dT27451 dT13837 dT5926 dT16527 dT15344 dT11506 dC_15179 dT25711 C7_118131103 dT21395 20 dT18480 dT50763 dT18102 dT8126 C1_220011116 dT3773 dT38048 dT50366 S4_17912300 C4_182430864 dT5186 dC_15863 TS5_128053793 dT13827 dC_388382 dT36120 dT50264 dT22636 dT7396 dT33944 C2_22518425 dT30922 dT50307 dT8749 dT8512 C5_156998414 C5_112498054 dT49047 dT50419 dT18762 dT29470 dT9283 dT1873 dT48390 dT12488 dC_10270 C3_216897716 dC_24267 dT24186 dT3360 dT34497 dT32431 dT40427 dT14669 dT3402 dT32245 dT649 dT37132 25 dT27345 dT30788 dT24408 dT50697 dT9029 dT8327 dT46843 dT5081 dT53247 dT28896 dT34184 dT9315 TS7_180945596 dT26006 dT50826 dT22681 dT31449 dT23515 dT45821 dT53022 dT12194 dT47840 dT17674 S5_40839446 dT46209 dT15569 dT47548 dT44178 dT49212 dT13764 dT52847 dT48530 S7_39646339 dT5917 dT14387 dT4737 dT11877 dT37911 dT45924 dT7985 dT17379 C2_21121892 dT28837 S7_39646309 S3_296004281 dT1832 dT19812 dT31217 S2_110636956 dT12456 dT28906 dT741 30 dT24786 dT732 dT18515 dT38881 S7_39646292 dT50258 dT32027 dT16449 dT38026 C1_102656250 dT9337 dT51420 dT33677 dT16793 dT12924 dT13338 dT46427 dT36036 dT3510 dT35551 S7_133861267 dT44856 dT21645 dT13498 dT34879 dT54030 dC_79210 dT19473 dT32695 dT29465 dT19298 dC_79234 dT11200 dT37277 dT8557 C6_223591048 dT33261 dT4836 dT50794 dT18485 S3_234961343 dT43636 dT49950 dT1762 dT1355 35 dT1848 dT14658 dT1593 dT15678 dT21114 dT46948 dT31735 S2_76988086 dT41196 dT43106 dT23823 dT26926 dT3554 dT42109 dT51201 TS2_76988086 dT30018 dT46299 dT49039 dC_183703 dT40903 dT23838 dT50145 S2_76982335 S7_100645462 dT47400 dC_1926 dT40278 dT32011 dT17535 dT23182 dT23364 dT12951 dT5796 dT21265 dC_9578 dC_4103 S7_83881022 40 dT43048 S4_49830629 dT23237 dT19939 dT50180 dT41244 dC_4097 dT7491 dT31151 C5_132196063 dT16824 dC_23097 S4_53822034 dT28498 S6_209490431 dT11408 dC_9584 dT38871 dT23798 S3_237947843 S3_65392670 dT20788 dT17391 C4_54042612 dC_15219 dT51779 S6_199487878 dT43186 S2_78894878 C4_136757390 dT29364 dT15747 dT32207 dT20397 dT20953 dT43458 C6_200342230 S2_78894944 dT47650 dT15626 dT24792 dT50336 45 dT6993 dT35908 dT35845 dT36645 dT34779 dT5625 dT29597 dT27602 S6_197918666 dT46652 dT35239 dT10689 dT14825 dT5196 dT19647 dT8425 dT28312 dT23469 dT14180 dT25817 dT30235 dT14780 dT22667 dT48800 dT27717 dT10549 50 dT15391 dT9008 dT52869 dT11768 dT12193 dT3431 dT4979 S2_78894908 dT46085 dT8910 dT1524 dT53077 dT8566 dT48497 dT6688 dT22956 dT10507 TS5_23967052 S1_17257463 dT14375 dT15063 dT11865 dC_42042 dT31410 dT22796 dT50903 S5_34147564 dT9167 dT2147 dT348 dT36870 dT48992 dT22841 55 dT31722 dT5877 dT51949 dT28049 S3_68670889 dT34541 C2_206875269 S1_135580638 dT48243 dC_28051 dT5016 dT45884 dT40868 dT47714 dT46811 S1_17257559 dT22363 S2_192440020 dT28517 dT10571 S1_17257551 dT18219 dC_212370 S3_68670908 dT48482 dC_223673 S3_255053440 60 dT18424 dT33771 dT36949 dT51287 dT45174 dT973 dT10378 dT47961 S3_255053465 dT51194 dT7119 dT5487 S3_257143847 dT8459 dT13022 dT46228 S4_14538257 dT1149 TS1_248242945 dT43632 dT33654 dT11302 dT34571 dT35964 dT24276 dT33655 dT37165 dT19739 dT48112 dT31033 dT3675 dT17340 65 dT40851 dT18867 dT3390 dT44398 dT53672 dT45970 S7_213184560 dT6729 dC_399018 dC_45605 dT22957 dT12272 dT11544 dT2290 dT22858 S3_253954916 dT19245 dT8998 dC_359278 dT42647 dT3272 S6_9770308 dT32036 dT29965 dT3906 dT3928 C1_133709193 dT42768 dT10652 dT37501 dT29409 dT28463 dT45243 dT4427 dC_60915 dT24570 TS5_12313000 70 dC_339765 dT50758 dT44368 dT48835 dC_107792 dT1352 dT10696 S1_171369178 dT6927 dT9233 dT19025 dT51291 TS4_14538257 dT2536 dT40041 S3_92611362 dC_167176 dT7711 dC_1015698 dT7856 dC_10879 dT44581 dT7910 dT18064 dT40470 dT37158 dT3180 dT1549 C4_99397060 dC_90916 dT39297 dT971 dT31571 dT12246 dT44853 S3_244121865 75 dC_31636 dT17835 dT22725 dT45101 dT31485 dT25838 dT29691 dT5489 dT51677 dT44636 dT10843 dT19448 dT50977 dT18863 dC_64162 dT14945 dT40635 dT51116 dT43892 dT2317 dT32308 dT47132 dT47057 dT53887 dT18489 dT51942 dT12496 TS4_171269213 dT17639 dT50618 dT42995 C1_234266014 dT50932 dT35462 dT30560 dT41167 80 dT30722 dT4343 dT51615 dT47137 TS1_273230575 dT25232 dT19043 dT27325 dT33642 dT29654 S2_236302464 dT30844 dT32392 dT11234 dC_88689 S3_276917001 dT48559 dT12209 dT35127 dT45124 S5_7812618 dT7314 dT44224 dT26079 dT48120 dT12709 TS4_171269280 dT6817 dT24369 dT766 dT42117 dT16041 S2_246378092 S7_182689743 C4_194142272 C4_194142277 dT15991 dT33254 85 dT36979 dC_193114 dT13169 dT32534 S7_182689737 dT8377 dT1085 dT48401 dT10533 S1_217117736 dT18155 dT25285 dT6193 S7_182689721 dT16658 dT18211 dT9213 dT48984 dT30297 dT20949 S7_90744125 dT19423 dT51346 dT49009 TS6_16445661 dT40084 dT3813 dT12622 dT28918 dT51331 dT35381 dT28196 dT47034 dT23737 dT46712 90 S1_241274562 dT50361 S7_201830035 dT23909 dT10670 dT6590 dT34102 dT52173 dT37923 dT31503 dT10131 dT30367 dT15893 dC_23232 dT50718 dT42534 dT6207 S1_268164908 dT31430 dT429 dT23052 dT47482 dT35323 dT22708 TS6_395354 dT37270 dT43605 S7_32127059 dT2763 dT16487 dT27517 dT49410 dT26411 C8_45958102 dT13489 dT33050 dT40596 95 dT16409 S3_35602619 dT52149 dT42227 dT27051 dT33550 S7_117332098 dT48669 dT30275 dT42183 dT4135 dT3175 C6_73259826 dT31375 dT10926 dT36791 dT37207 dT35318 dT35742 C6_59409932 dT24125 dT36311 dT37626 dT7088 dT7307 dT31619 dT16616 dT43871 C6_59409930 dT33782 dT37138 dT526 dT21724 dT39129 dT16875 dT2820 dT32206 dT771 dT46112 100 dT37696 dT30134 dT8609 dT17908 dT492 dT6664 dT1888 dT40500 dT40810 dT2179 dT35232 dT6698 dT35582 dT1427 dT47247 dT37914 dT45831 dT53184 dT30790 dT15707 dT12717 dT8114 dT12621 TS7_167047574 dT49612 dT34207 dT49861 dT32542 dT53440 dT26512 105 dT52369 dT32848 dT50319 dT42678 dT40421 dT49424 S7_102687604 C7_89265590 dT40659 dT22787 dT18556 dT14806 dT29318 C4_41131304 dC_221873 dT22104 dT18945 dT4190 C7_82320373 dT14307 dT22647 dT29815 dT40952 dT50943 dT49133 C3_245369357 S5_31904873 dT38935 dT3713 dT46312 110 dT32996 dT43387 dT39799 dT31316 dT22267 dT15879 dT43702 dT3053 dT358 dT12440 dT8813 TS3_205733261 S5_26740824 dT45995 dT28003 dT32512 S6_398797 dT24042 dT4924 dT37762 S8_44730456 115 dT52226 dT48403 dT26721 dT15989 dT26387 dT23856 dT6733 dT17794 120 dT42771 dT207 S5_16898738 dT29058 dT21277 dT37016 dT46613 dT38486 dT47995 C3_202051529 dT34310 125 dT8152 dT51527 S7_89265546 dT15649 TS5_13509703 dT15821 dT54119 dT53202 dT7678 dT22318 130 dT31754 dT32125 S4_86603238 dT3519 dT27794 dT25849 dT2435 dT42236 dT35984 dT46426 dT20384 135 dT23371 dT25577 C5_3132406 dT1466 S3_41221675 dT15117 dT29259 dT780 140 dT40376 dT15952 dT26168 dT36121 dT32758 145 dT30470 dT42316 dT53493 150 dT3930

155

160

165 dT159 dT590 dT6548 170 dT41013 dT38488 dT6810

175 dT50779 dT17665

Figure 2-7. Genetic linkage map of the napiergrass female parent N190.

Napiergrass: Male Parent Linkage Map

1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 S3_66915283 dT11902 S3_260011343 S5_3204810 S2_14223558 S4_192354592 S2_174467255 S1_256402076 dT29596 dT25323 dT47338 dT52537 dT36565 S1_78149353 S3_66915258 dT17348 S5_3204852 S5_3204827 dT30464 dT35994 dC_4861 S1_78149366 dT6838 dT46533 S5_3204849 S2_16747159 dT6813 C5_3132373 dT43261 dT16456 dT4830 dT44555 dT53963 S5_3204807 dT8117 dT8719 dT26692 dT37861 dT25275 5 dT27134 dT38576 dT28731 dT3062 dT33885 dT43606 S5_2055822 dT18503 dT31744 dT41707 S3_66915278 C3_124558897 dT41513 dT45868 C2_14084192 C8_42747377 dT40955 dT10276 S1_7068544 dT10674 dT12415 dT20572 dT26674 TS7_92021240 dT25964 S3_43511354 dT46296 dT14701 dT32396 dT3437 TS6_10754586 dT15685 dT9434 dT44870 dT13178 dT39601 dT39911 10 dT30975 dT46011 dT10631 C2_174576603 dT19710 C5_1635981 dT52605 dT38476 dT27467 dT8760 dT1897 dT13445 dT51406 S1_203215691 dC_94047 dT35347 dT2555 dT16974 dT53530 dT8898 dT31069 dT51571 dT50501 C2_174467255 dC_58121 dT12412 dT33533 dT45550 dT23430 dT29388 dT24568 dT27547 15 dC_58157 dT32705 dT28618 dT11037 C5_4168325 dT7310 dT7731 dT46387 dT8606 dT43154 S3_248431947 dT13383 dT1325 dT9771 dT39822 S5_31045825 dT21831 dT48623 dT4991 dT36424 dT38592 dT34088 dT25225 dT24339 dT28250 S6_35473070 dT36067 dT19877 dT45191 dT33344 dT41728 dT25619 dT33835 20 C5_27253496 dT35938 dT50761 dT23648 dT14502 dT4180 dT25168 dT31852 dT15831 dT10459 dT11256 dT2397 TS7_74798545 C6_39611566 dT941 S2_30824637 dT40880 dT39395 dT44314 dT48139 dT52160 dT29846 dT21208 dT42952 dT41528 dT14455 dT31263 dT6835 dT45539 dT5478 dT25394 dT53592 dT9903 dT20879 dT3030 dT41386 C4_131830009 dT11071 dT6267 dT16233 dT22371 dT22002 dT32461 dT47535 dT9904 S6_32125800 25 dT27514 dT16455 dT36198 dT52290 dT30050 dT38911 dT32024 dT51889 dT1174 dT5307 dT9078 dT40535 dT18538 dT38126 dT40798 S8_31103785 dT20355 dT32166 C1_119093525 C1_119093531 dT38805 dT3345 dT42630 dT2510 dT20991 dT1844 dT43 dT48489 dT20154 dT3073 dT47937 dT11671 dT37316 dT6794 dT1791 C4_200698121 dT4388 dT47375 dT24760 30 dT34717 dT24498 dT19937 dT53836 dT12054 dT14931 dT32449 dT4430 dT8297 dT23509 dT48033 S1_71014695 dT16992 dT28199 dT37466 dT13687 dT23119 dT15729 dT38885 dT36533 dT13802 dT17414 dT46041 dT41204 dT43805 dT33449 C2_38440300 dT43027 dT29057 dT16346 S5_15050945 dT32273 dT48964 dT30022 dT6561 35 dT3891 dT28736 dT22920 dT52180 dT47931 S3_284635389 dT47787 dT51314 dC_13461 dT45351 dT9367 dT24969 dT7886 dT2475 dT42989 dT6913 dT1264 dT43643 S4_116088663 dT50610 dT36410 dT31904 dT35753 dT18370 dT12390 dT50566 dT13912 dT11516 dT1444 dT13255 dT52321 dT11891 dT16483 40 dT41113 dT9987 dT15477 dT32416 S6_161712969 dT29644 dC_121093 C5_13708967 dT30582 S5_32287078 dT11178 dT3792 dT12612 TS4_61117193 TS4_175268085 dT53121 dT29726 C4_90405411 dT44347 dT36437 dT5999 dT9304 dT16762 dT22293 dC_618193 C2_69640554 dT777 dT8320 dT6436 dT13560 dT5577 dT1436 dT44577 dT8400 dT42843 dT6043 dT6248 dT23975 S4_29656682 dT4353 dT29458 45 dT34797 dT34457 dT1869 dT29868 S3_123721011 dT35828 dT31396 dT14717 dT35978 dT42579 dT18940 dT39279 dT22493 dT51180 dT3108 dT4620 dT4379 dT5545 dT49798 dT11704 dT43235 dT5027 dT13681 dT28611 dT24372 dT13121 dT1631 dT2661 dT45689 dT304 dT18120 dT2439 dT23432 dT2286 dT30154 dT35867 dT4303 C4_54525188 dT24621 C5_23963287 dT17003 50 dT29788 dT39900 dT16070 dT23729 dT1297 dT27692 dT52876 dT30035 dT52425 dT6792 dT39167 dT42777 dT43233 dT861 dT47337 dT30575 dT34784 dT6238 dT15502 dT18050 dT45026 dT24471 dT18870 dT23905 dC_161538 dT37028 dT43507 dT20275 dT43430 dT18206 dT51168 dT34144 dT8454 dT42142 dT25258 55 S1_193475897 dT34243 dT11454 dT14389 dT1961 dT31045 dT36388 dT53270 dT7751 S7_227510 S5_15660406 dT38621 dT35142 dT12204 TS3_259950719 dT3762 dT9893 S7_161179045 S7_161179072 dT27267 S1_211915643 C5_64002238 dT991 dT47792 dT38724 TS3_102980429 dT27178 S2_187091485 dT8498 dT12112 TS3_71103541 60 TS1_74139971 dT11106 dT5958 dT24088 dT21588 dT33297 dT5896 dT19279 dT1716 dT44755 dT34857 dT6132 dT8821 S3_277598650 dT5137 dT3606 dT14412 dT29230 dT53659 dT26322 dT48904 dT7984 dT3273 dT24075 dC_37024 dT8867 dT16 dT4496 dT14775 dC_66514 dT30639 dT23286 dT48523 dT13772 G3_243475604 S3_243475604 dT19864 dT13934 dT21669 dT48116 65 dT6466 dT40485 dT38411 dT10105 dT53762 dT25121 S3_243475658 S5_77555777 dT47510 dT45910 dT46519 dT16151 dT34305 dT53712 dT41621 dT45471 dT47317 dT8946 dT39882 dT36593 dT4800 dT50561 dT10955 TS4_49707946 dT32613 dT38599 dT38317 dT18605 dT30721 dT19708 dT50908 C3_273261021 S5_52320341 dT743 dT49164 dT8487 dT46588 dT11273 dT5751 dT51740 dT12000 dT40493 70 dT29128 dT36233 S3_243475784 S2_151241666 dT5755 dT43759 C2_186899428 dT46261 G3_47707669 dT43569 dT52531 dT18641 dC_75778 C2_151270539 dT46787 S4_109511769 S6_69712882 dT23775 S3_47707650 dT41949 dT28669 C2_168637894 dT52704 dT49723 S5_59674699 TS4_79794887 S6_69712890 dT9800 G2_236219679 S3_258304119 dT47752 dT15498 dC_1941 dT41484 dT39328 dT44455 dT33950 75 dT36845 dT41069 dT32303 dT27423 S3_258304047 S3_258304050 dC_5486367 dC_428936 dT6670 dT45195 dT48885 dT22005 dT33420 dT31386 S2_15995145 S2_15995173 dT35339 S3_258304063 dT50064 dT41460 dT17358 dT2081 dT26938 dT40660 dT28525 dT25340 dT17613 dT9518 S2_15995166 dT30147 dT40541 dT5222 dT40996 dT44451 dT7562 dC_124707 dT51317 dT25206 S3_50037874 80 dT43295 dT7695 dT31232 dT14352 dT22374 dT31777 dT20193 dT31778 dT25091 dT31359 dC_242221 dT2126 dT34738 dT29968 dT24215 S4_17921468 dT3770 dT10587 dT31173 dT50771 dT44933 dC_150274 dT52854 dT29452 dT32675 dT39406 dT7877 dT51103 dC_198183 dT32403 dT29828 dT33834 dT17019 dT11711 dT40958 85 dT27185 dT48427 dT45232 dT26827 dT34859 dT29535 dT14016 dT46138 dT43835 dT35398 dT10167 dC_26149 dT51146 dT4301 C4_13275000 dT29908 dT30984 dT5445 dT17380 dT33063 dT2668 dT22682 dT23969 dT24376 dT34795 dT7580 C3_198785997 dT52592 dT35972 dT7992 dT8511 dT10481 dT36267 dT48051 dT25812 dT574 dT24195 dT27463 dT22362 dT34945 dT49967 90 dT7381 dT26600 dT53110 dT11183 dT53719 dT987 dT5635 dT24964 S7_168784407 dT27669 dT17028 dT13083 S4_21013305 dT52456 dT16879 dT35258 dT14813 dT17187 S3_267455176 dT20846 S7_168784355 dT16149 dT43713 dT10193 C5_27253540 dT9307 dT10669 dT11551 C3_119371432 C2_56055562 S2_191695057 dT28164 dT39674 dT34107 C6_186956079 dT47900 dT21093 dT40850 95 S3_12254914 S2_191695048 dT838 dT1564 dT3299 dT22232 dT37205 dC_299275 dT6981 dT2704 dT49783 dT34453 dT21673 dT29166 dT9256 dT12426 dT51487 dT35529 dT32941 dT24290 dT49964 dT47823 dT48272 dT2229 dT12898 S4_17796331 dT26372 dT18627 dT334 dT27447 dT35854 dT32474 dT5924 dT10535 dT1781 100 dT34136 dT53079 dT32069 dT50663 dT46240 dT10724 dT49429 dT2057 dT16697 dT29456 dT52542 S7_12814263 dT2770 dT28451 dT1963 S2_116787266 dT43200 dT18564 dT11525 dT47089 dT36456 dT45454 dT221 dT13262 105 dT6453 dT51978 dT39089 dT11880 dT53771 dT51060 dT38416 S2_225137713 dT32197 dT19151 S6_99869597 S1_49814862 dT36711 dT48311 dT8342 dT52360 dT47043 S2_201762543 dT46410 dT22571 S6_99869584 S2_225245713 dT52029 dT25766 dT52469 dT33021 S7_146917990 dT30920 dT36501 dT9432 dT1280 dT32617 dT48485 dT34492 110 dT37356 dT1490 dT31921 dT11509 TS3_132668864 dT3322 S2_223315521 dT29482 dT14203 dT35128 dT51442 S7_146917952 dT4155 dT15069 dT8391 dT16591 S5_136028687 dT53136 dT26122 dT15629 dT1086 dT40139 dT52684 dT38040 dT47563 dT39139 dT35797 S7_100737594 dT41891 115 dT20499 dT41347 dT22927 dT30730 C3_64621327 S5_132636460 S5_132636448 dT36338 dT21823 dT35508 dT20567 dT11606 S5_132636508 dT723 S2_214316973 dT33180 dT27310 dT19693 S6_118469241 dT38799 dT9509 dT332 dT23060 dT30544 dT10485 dT46088 dT35795 dT44366 dT14954 dT27949 dC_45050 dT28689 dT9291 dT41539 120 dT28974 dT39428 C5_65406593 dT36827 dT24014 dT45475 dT33330 S1_249594415 dT19912 dT49533 dT36253 dT41922 S5_46066941 dT1374 dT5535 dT53574 dT20066 dT11287 dC_247492 dT29657 dT14847 dT8792 dT43321 dC_282880 125 dT2450 C7_38012075 S7_52134705 TS2_146426838 dT20514 dT42596 dT34936 S2_100034039 S7_52134761 dT31286 dT48305 dT2297 dT38859 S5_142555945 dT14670 dT17087 dT32706 dT8340 dT41084 S7_52134764 dT38863 dT43816 130 dT49414 dC_234325 dT47758 dT38020 dT10476 C1_251517594 dT20274 dT9747 dT5281 dT33605 S2_226559654 dT53934 dT24863 dC_259973 dT22921 dT20176 dT50383 S2_216176645 dT13782 S1_251483789 dT30440 dT7627 dT4281 dT9060 135 dT41404 dT14807 dT40123 dT45959 dT25565 dT25902 dT42403 dT30703 C5_154569850 dT32999 dT44700 dT43823 dT9863 dT6973 dT46507 dT20166 S2_176915862 S1_260924412 dC_196800 dT21342 dT717 dT40245 140 dT17024 S3_45974916 dT35453 dT18761 dT12737 dT49728 dT5816 dT10443 dT45297 dT36412 dT51649 dT10471 dT13896 dT28492 dT24998 dT48862 dT24540 dT17979 dT21007 145 dT47753 dT25533 C3_67469347 dT19286 dT17489 dT31973 dT24044 dT32523 dT29143 dT26209 dT43692 dT20483 dT26005 dT35466 dT21677 S6_55827481 dT15881 S5_1337832 dC_2450640 dT53506 dT23441 dT29640 150 dT23539 dT48962 dT36750 dT28159 dC_75137 dT50248 dT13412 dT9688 dC_162440 dT27136 dT43184 dT36706 dT35101 C6_200342258 dT32605 dT15816 dT11814 dT35136 155 dT9665 S1_267892930 dT5428 dT16180 dT36597 dT18516 S5_116583893 dT26922 dT50685 dT18539 S3_215627633 S5_116583902 S7_119895303 dT10621 dT13751 160 dT53635 dT8614 dT9181 dT50350 dT36969 dT29758 S7_216787106 dT22426 dC_45074 dT10112 165 S3_58384211 dT32609 S3_58384176 dT25438 dT17747 dT18781 170 dT26851 dT45130 dT35166 dT29238 TS3_52460829 dT44698 dT6553 175

dT13151 dT37830 S7_66630588 180 dT39337 dT44668 dT25362 dT31763 S3_61665185 dT28445 185 dT9471 dT3618

190

195

200

205 dT20803 dT13705

210 dT24779

Figure 2-8. Genetic linkage map of the napiergrass male parent N122.

Figure 2-9. Genotyping by sequencing single nucleotide polymorphism (GBS-SNP) marker distribution for the 14 linkage groups of napiergrass. A black bar means a GBS-SNP marker. A blue bar represents segregation distortion region. The left scale plate represents genetic distance (centiMorgan as unit).

Napiergrass: Integrated linkage map LG01 LG02 LG03 LG04 LG05 LG06 LG07 LG08 LG09 LG10 LG11 LG12 LG13 LG14

0 S1_267892930 S6_80432248 St_36922 S5_3204810 S2_30030493 dT42156 S7_219034180 St_62484 dT27092 dT37914 dT34863 dT52537 dT19752 S7_177490027 S1_241274562 dT46838 St_27049 S5_3204852 dT13183 dT36401 St_63298 dT46112 dT43981 dT16456 St_57083 dT36338 dT35970 St_27050 S5_3204827 dT41204 dT53206 S6_15199694 dT41558 dT32288 dT26692 dT6944 dT39139 dT13151 dT24472 S5_3204849 dT51314 dT12209 S6_37610339 dT40810 dT34945 dT43606 St_51502 dT53136 dT29238 S3_85079300 S5_3204807 dT6913 dT49934 dT45995 dT14190 dT33782 dT16149 C2_14084192 St_52583 dT35128 S6_197918666 dT11506 St_55016 dT11178 TS4_171269213 dT2684 dT36311 dT8511 C8_42747377 dS_53273 5 dT3813 dT17747 dT38048 dS_69164 dT8400 dT22615 dT31673 dT44709 dT31375 dT2668 dT14701 dS_30098 dT37356 dT18539 St_36360 St_83036 dT3792 dT2317 dS_29023 TS8_54685526 dT18082 dT49930 dT31069 dT47563 dS_22638 dT50350 St_36361 St_55233 C2_69640554 dT23827 dS_7808 dT32644 dT26938 St_83499 dT31151 dT52469 dS_64625 dT44178 S5_34645373 dT777 dS_49354 dT7834 dT12313 dS_15838 St_26088 dT9771 dT15747 dS_58585 St_65272 dT7396 St_56474 dT34457 dT27674 dT52932 dT27051 dS_48139 dT2397 dT14203 dT19814 10 St_7983 dS_36562 S3_234961343 dS_12036 dT13912 St_48501 dT1325 dT19776 C8_45958102 dT31173 dT41528 dT34137 dT24731 dT8342 dT22426 S3_65392670 St_56231 dT15477 dT51677 dC_58157 dT29318 dT35420 dT30424 dT11071 S5_136028687 dT51987 St_8341 dT11200 St_35372 dT8760 dT32273 dC_31636 dT24339 dT1374 dS_28564 dT18438 dS_70286 dT8126 dT21395 TS1_57392744 dT29974 dT35908 St_57688 dT3108 C4_99397060 dT33835 dT41539 St_43790 S4_131830044 dT20355 dS_23166 S3_258304063 dT30392 St_63794 St_31668 St_50301 dT28199 dC_90916 dT10459 dC_9265 dT40690 dT44002 dT47937 St_51337 S3_258304050 dT41386 S3_258304047 15 dT36979 St_63444 TS3_40995143 dS_3673 S2_16737601 S4_43915384 dT33330 dT37270 dT15502 dT47375 dT36120 dT766 dT2241 dT26209 dS_33240 dT1791 dT27183 dT21208 TS2_146426838 dT4159 dT52876 dT29957 dT51352 dT34497 dT36115 dT47054 dT2147 dS_55772 S2_16737644 S4_43915389 dT38126 dT32668 S1_268164908 St_72633 St_23648 dT18762 S3_258304119 dT7314 dT14807 dC_42042 St_56111 dT42630 TS4_199335191 dT20154 dS_68537 S1_268948076 dT45689 dT23509 dT3360 S3_237947843 TS1_242079107 dS_24618 S3_68670889 dT22629 dT51889 dT21313 S7_182881700 St_67003 dT9656 G4_131830067 dT13802 St_50213 dT11865 dT30844 dT30703 St_36963 C5_4168325 St_22811 dT30293 dT16455 dT25459 dT27080 dT29868 C2_38440300 dT12719 S3_259562492 20 TS1_273230575 dC_196800 S3_68670908 St_52503 dS_12677 dT47539 dT24498 St_67005 St_8337 S4_131830067 dT47931 dT34794 TS3_202192213 C1_234266014 dT24998 dT5016 dT36067 dT22002 St_40435 dT16860 St_69926 dS_20051 S4_131830066 dT9367 dT25766 dT22841 dT47057 dT30440 dC_28051 dT21831 St_23394 TS4_17374522 dT6436 St_8239 dS_5062 TS4_175268085 dT35753 dT34959 dT43569 dT19448 St_68017 dT22363 dT50761 S2_174624948 dT52456 dT4353 dT771 TS1_270056181 S4_131830063 dT16483 dT15762 dT40493 dT10843 dS_24309 dT33771 dT941 dT5178 dT23773 dS_40842 St_71627 dT23820 dT13560 dT29644 dT19372 dT27298 25 dT22725 dS_50465 dT51194 dT50341 dT43720 S4_17921468 dT39900 dT35795 dT52173 dT11891 dT36437 dT34492 dT27432 dT971 dT41470 St_27334 dT23430 dT28282 St_40403 dT23432 dT40500 dT23737 dT22876 dT5577 dT43443 dT47317 dT18064 dT8340 St_27337 St_52805 dT40697 dT45232 dT861 St_60317 dT40084 dT31904 dT54067 dT41039 TS3_52779446 dT37631 dS_78648 dT13022 St_54858 dT38918 dS_43008 dT23905 St_65388 dT48984 dT12269 dT36404 dT45156 TS3_52779443 S7_227510 dT48523 dT40041 dT2297 dT40851 dS_40368 S2_24418753 dT39406 dT33180 dT18940 dT10533 dC_55022 dT29458 dS_63438 dT30583 dT10571 S1_171369178 dT41082 dT42647 St_54860 dT12948 dS_62092 St_65387 S5_74991349 dT33254 dT22567 dT37862 dT44382 30 dT42142 dT10105 dC_339765 dS_20475 dT42768 dT41977 S2_24418742 dS_59686 dS_43802 dT30575 dT26907 dT38883 dT5545 St_57235 S7_161179045 dT48904 C1_133709193 dT42484 St_26265 dT22371 dT42842 St_44404 dT52684 dT47221 St_7976 dT28123 dT304 dT20917 S7_161179072 TS3_71103541 dC_359278 S6_242118514 TS3_109270808 St_53749 dT19311 St_40987 dT47247 dT38621 St_7494 dT26116 dT24621 dS_5311 S7_112269 dT43079 dT13670 dT40470 dT50758 dS_24959 St_16460 S4_56719430 C6_59409930 dT9800 dT24369 dT10040 St_18279 dT7338 G7_212623362 dT47792 TS1_149192348 dT31260 dT6927 St_53600 dT38665 dT12456 dT24125 dT45548 dS_35489 dS_61343 dT17058 dT48311 35 dT46230 S3_255053440 dT45509 dT42554 St_27574 dT36706 dT33063 dS_16097 C6_59409932 dT36456 S2_236302464 dT36340 dT30035 dT33358 dT34305 dT4657 dT12272 dT45101 St_26191 St_53555 dT47411 dT49039 C6_73259826 dT2536 dT45124 dT26152 dT18050 dT11880 dT910 S3_255053465 dT53672 dT50977 St_27576 dT31058 St_7438 dT9886 dT47034 dT9428 dT20774 dT10739 dT51168 dS_66625 dT14775 S3_257143847 dT48112 dT25232 dT24195 dT13412 St_33827 dT1848 dS_59431 dT1352 dT47132 dT26884 St_19470 St_68482 dT26620 dS_45674 dT33654 dT15707 TS3_34404935 dT36750 St_33826 St_39084 dT1281 dT18564 dT51615 dT621 dT14389 dS_4586 dT43326 dT35964 40 dT8459 St_52312 S3_92611362 dT53506 dT3578 dT38946 TS6_101968558 dT24570 dT50618 dT17578 dT1961 St_52257 dT27451 dT17340 S1_144792807 dT13790 dT52592 dS_64191 dT50366 TS4_56235778 dS_15188 dT1670 dT43193 St_80208 dT12204 dS_46016 dT50763 dT6792 dT973 dC_193114 dT12246 dT26005 dT33944 St_47561 dT454 dT7119 S1_256031621 S7_89250999 dT27243 dT28451 dC_388382 dT17003 dT16562 dT42117 dT24290 dT17979 dT1873 St_47563 dT17688 dT5487 dT47532 dT17325 dT991 dT15626 dT11554 dC_399018 dT18424 dT2283 dT2704 dT717 dT14669 dT35551 dT22268 dT50663 TS1_55816576 dT25711 dT12112 dT16697 dT24186 St_32898 St_2847 dT11665 St_27619 dT32999 dT8327 TS4_79794887 TS6_16445661 dT44398 dT144 dS_21720 dT29731 dS_12408 45 dT50419 S3_253954916 dT50103 St_67496 St_27618 St_54089 St_17032 dT28999 C6_105814832 dT3390 dT40320 St_16173 dT19317 dT16760 dT24408 dS_3349 dT28049 dT16875 S3_12254914 C5_154569850 dT31449 St_47706 dT29814 dT48482 dT6229 dS_353 dT44755 dS_26459 S7_162635983 dS_1370 dS_5480 dS_65719 dT31485 dT20176 dT13764 TS4_79798648 dT52515 dT37165 dC_220532 dS_12928 dT37721 dS_49110 dT46209 dT29965 S1_17257551 St_70712 St_27764 dT15952 dT49212 TS4_79798642 dT31219 dT10724 dT50746 dT5186 dT26322 dT14825 S7_100737594 dS_33295 St_1775 dT4135 dS_63784 dT7627 S2_110636956 St_47414 dT13258 dT1149 dT45573 dT8749 dT39135 dT13865 S7_133861267 dT11704 50 St_1908 dS_32733 St_28126 dT780 C2_21121892 dT33420 dT52774 TS1_248242945 dS_30445 dC_10270 dT23286 dT22667 dT8557 dT45243 St_3838 St_69789 dT20803 dT9747 dS_26130 dC_124707 dT9518 dT8998 dT7085 dT15950 dT38411 dT12193 S7_52134764 dT48835 S1_17257559 dT18264 dT22019 dT47758 St_24033 C3_186584463 dT32504 dT22957 dT20685 TS8_38834569 dT45471 dS_63968 dT36827 dT35828 dS_38214 St_12035 dT18489 S5_92980877 dS_49336 dT37277 dT38233 dT3928 dT17967 dT649 dT38599 dT8566 dT46088 dT6248 dS_33009 dT35582 dT18863 C5_3132406 dT16449 St_47413 St_45444 dT45174 dT6028 dT53247 dT12000 St_54452 dT14658 dT22293 55 dT48243 St_864 dT37926 St_57085 St_41249 dT30894 St_60942 dT26839 dT36552 dT53022 dT4552 TS5_23967052 C7_146983287 dS_335 S1_17257463 dT35318 S3_99632078 dT15117 dT36036 dT9337 dT4343 dT21093 dT49336 dT12194 dS_3981 dT36565 dT36253 dC_1015698 dT48497 dT16616 dT13705 dT46426 St_12768 dT36845 S6_59035298 dC_299275 dT23044 S7_39646339 G3_47707669 S5_34147564 S7_100645462 dT1549 dT3431 St_58109 dS_55508 dT20384 dT29465 dT18515 S6_199704613 dT28517 dS_59001 S7_39646309 dS_27715 dT51949 C7_38012075 S3_244121865 dT48800 dS_35557 dT26079 St_55733 dT49950 dT52450 S6_241915017 dT40868 dT47746 S7_39646292 dS_68831 S5_875176 dT25902 dT29691 dT14180 dT16409 dT31367 St_55732 St_18051 TS4_114218088 St_62949 St_13976 dT29658 dT33677 St_28245 dC_4861 60 dT43823 dT51116 dT42594 dS_66127 dT16041 dS_40895 St_13306 St_32476 S6_241915045 C2_206875269 dC_105219 dS_17159 St_28242 C5_3132373 dT37501 dT12496 C1_212950043 TS8_83771289 dT42473 dS_35332 St_75341 S4_55762918 St_59677 dT36870 dT2822 dT3510 dT26266 TS5_12313000 dT43692 dT41167 TS1_134040623 dT35381 dS_37174 St_54510 S2_76988086 dT10273 dT45221 dT334 dT39692 dS_36568 S3_47707650 S5_2055822 dT19286 dT33642 dT46801 dC_23232 dT13169 St_57106 TS2_76988086 S4_71863256 dT23999 S2_78894908 dT40261 dS_18627 dT9730 dT10696 S5_1337832 S3_276917001 dT29597 dT47482 dC_150274 dC_282880 S2_76982335 dT52847 dT7731 St_58417 C1_149814711 dT19298 dS_58321 dT51406 dT14025 dT12709 65 dT6993 dT10670 St_33851 dT10717 dT5796 dT5808 dT4991 dS_34607 dT31390 St_35836 dT13363 dT51571 S7_90744125 S7_182689721 C1_117276997 dT16487 S3_37058907 St_53813 S2_14223558 St_46189 St_19657 dS_44017 dT43343 dT1762 dT27423 dT34088 dT28159 dT317 dT36356 dT16658 St_30077 dT20066 dS_3168 dT30598 St_67385 dT52869 S1_235015034 dT41196 dS_2736 dS_26201 dT9688 dT9078 dT38584 dT22201 St_35238 dT27794 dT50180 S4_109511769 dT45191 dS_13200 S1_235015054 St_76782 dT38555 dT41728 dT47365 dT19423 St_4264 dT21000 St_33736 C5_65406593 dT16824 St_46150 dT14502 C2_143807176 dT42355 dT30018 dT24050 dT25168 dT23441 dS_45055 St_4265 S7_154422455 TS3_274563801 dT9509 St_14790 dT11273 dT40880 dS_24485 dT12855 dT1053 St_26083 S5_7812618 70 dT11814 TS6_161742923 dT28498 dT7220 TS3_274563794 dT21823 dT1183 dT16151 dS_49183 dT36645 dT34010 dT40278 dT35339 dS_29656 dT26922 dT51331 dT13814 dT15893 TS3_274563751 dT32419 dS_68166 dT6466 dT4808 dT191 dT5814 S7_83881022 dC_15863 dT48033 dT16180 dT23909 St_4395 dT661 St_31749 S5_65361310 dS_22487 S4_143679490 St_64847 dS_13845 dT41014 dS_4342 dT3341 dT40635 dT10131 dT36064 dT23237 dT41813 dT24779 St_56639 dT1114 St_54552 dT47341 dT50464 dT30197 dT7491 C7_118131103 dT9904 S7_32127059 dT30367 dT13772 dT7888 TS3_123337609 dT2435 dT17391 dT8454 dT5478 dS_35262 dT5876 St_81113 dT40541 dT9903 75 dT36791 dS_40112 dT23364 dT38698 St_12442 dT50835 dT20397 dT18870 dS_30976 dS_13192 St_3875 S4_175268085 S3_146735713 dT44314 S7_119895303 St_31374 dT53762 dT42580 dT40996 dT15164 dT50336 dT43233 dT39816 dT51911 dT12762 dS_51373 S3_50037874 dT3855 dT33050 TS3_228287887 dT23838 dT5181 TS3_219893756 dT48427 dT27602 dT4620 dT32461 S2_225137713 St_9029 S4_175268116 dT25091 dS_21070 dT8614 dT7231 dT51201 dT50319 dT50718 dT5772 dT19647 dT42579 St_67353 S1_49814862 St_9025 dT32609 dT5917 St_77252 dT7088 dS_37949 dT45924 St_45778 dT7695 dT23323 dS_66177 dS_55849 TS6_200342427 dT40583 dT49435 dT9181 dT28896 dT46712 dT526 dS_36977 80 dT7984 St_30560 dT28196 dT30226 dT23469 S4_29656682 TS6_200143111 St_16188 C1_231986176 dT10621 dT37132 TS5_6700729 dT30134 dS_27204 dT46948 dT5239 dS_67157 dT8428 dT10549 dT3285 dT46758 dS_64130 dT40418 St_83873 dT8512 St_57050 dT35232 dT2763 TS1_74139971 dT39882 St_38576 TS5_114098156 dT4979 dT21909 dT47729 S2_225245713 dS_54187 dT29364 dT50771 St_56490 dT45831 dT40596 dT11106 dT30721 St_27151 S5_40839446 S2_214301707 dT8320 St_63770 dT17835 dT20183 dT20560 dT29828 dS_15609 dT53184 dT37207 St_11047 dC_100348 dT50908 dT41901 dT53077 C4_90405411 St_65266 St_33671 dT48992 dT35101 dT47840 St_59001 dT29758 dT8609 dS_83781 C2_129356854 dS_40283 dT40995 dT15063 St_41693 St_7565 dS_57503 dS_8637 dT37141 dT14016 St_77253 85 TS7_167047574 dT7307 dT19473 dT4190 dT22708 dT34274 dT348 dT32416 dT51264 St_21261 dT50903 dT35136 C3_216897716 dT30275 dT52369 dT31619 St_119 C7_82320373 dT5249 dT28525 dS_7312 dT13255 dS_53085 dT44636 dT10507 dT49802 dT38857 dT53121 S7_102687604 dT21724 dS_25631 dT22787 St_32104 dT48365 St_14914 dS_39004 dT330 dT32617 dT46085 dT29640 dT5445 dC_121093 C7_89265590 dT39129 dT732 St_38861 dS_2204 dT18480 dT34541 St_42645 S6_209490431 dT32308 dT11768 dT42825 dT19291 C5_13708967 C4_41131304 dT6698 dT13338 dT28192 dT6670 dT9315 St_15285 St_13970 dT15729 dS_23912 dT14780 dC_162440 C3_198785997 dT1436 90 dC_221873 dT30790 C1_102656250 dT36233 dT41484 St_57290 St_15284 dT29340 dS_31075 St_24952 dT28312 dT43565 dT574 dT49612 dT40952 dT42183 dT24786 dT31759 S3_35602619 S5_63556385 dT45884 dS_38178 dS_7074 S2_144268311 St_9993 dC_75137 dT53719 dT50943 dT10112 dT12412 dT7985 St_68287 S3_243475658 dS_51396 S2_216176645 St_48484 dT33449 dS_56758 St_8983 dS_53783 dT14813 dT22104 dT3713 dT34207 St_11044 dC_428936 dT1934 TS5_128053793 S2_192440020 St_31330 St_63375 S2_144268276 dT47682 St_116030 dT11551 dT46312 dT22267 dT32705 dS_9398 dT28669 S3_243475784 dT13827 St_16949 St_31333 St_70658 dS_14880 dT10689 dS_89120 dT24476 dT12440 dT40659 95 dT39765 dT5222 C2_168637894 dC_24267 St_18971 St_42445 dT26926 dS_57817 St_45245 St_116029 S3_58732195 S7_66705818 dT52605 dT18485 dT37647 dC_75778 C5_112498054 dC_212370 St_56407 dT8297 dS_41244 dT34779 dT50685 dT4006 dT32512 dT18945 dT47548 dT43295 C3_273261021 C5_156998414 dS_17901 dS_59418 dT1593 St_25765 dT47650 dT43419 dT40850 dT47732 St_30603 dT9029 dC_193851 St_34784 St_51734 dT51287 St_42444 C6_223591048 dT40973 C4_136757390 dT53109 dT34879 S8_44730456 dT27593 dT50826 dT19708 dT43871 S5_63556383 dS_50616 dC_13461 dT44856 dT29502 dC_9584 dS_217 dT6981 S7_89265546 dT39911 dT32431 dT42197 dT36593 dT27345 dT10378 dT32924 dT28906 dS_30742 dC_4097 dT10058 dT32941 100 S7_136957144 S4_86603238 dT29470 dT49133 dT2820 dT30788 dT46228 St_47777 dT11877 dT14696 dC_4103 dT31974 dT27447 dS_31623 dT50264 St_66871 G3_243475604 dT14651 dT11302 dS_57611 dT23515 dT20326 dC_9578 dT53392 dT50794 St_32380 St_5341 dS_16944 S3_243475604 dT15817 dT31033 St_47778 dT46843 dT35127 dC_1926 dT42256 dT10535 St_203 St_68737 dT492 St_50448 dT9291 dS_25947 dT18102 dT52863 dT40903 dT24540 dT2057 dT16527 dS_66207 dS_51063 dS_62152 dT45970 dT43805 dT13837 dS_47193 dS_6690 dT34174 dT9107 105 St_202 dS_19764 dT3273 St_49453 St_19647 dS_29894 dT15691 St_17971 dT1078 dT47569 dT2770 dT33481 dT46586 dT35742 dT4737 dT2290 dT23119 dT47337 St_17970 TS4_5133343 St_75947 dT6635 dS_62356 dS_50880 dT4631 St_50451 St_15938 C4_200698121 dT5999 dS_68589 dT42109 dT952 dT39970 dS_8179 dS_50638 S3_277598650 dT22005 dT3272 dT27514 dC_93403 St_20466 dT21114 dT54084 St_37822 dT46699 St_63927 dT12717 dT16773 dT10652 dT3030 dC_63791 St_20650 dT4836 TS7_195845045 dT221 110 dT52420 St_63925 dS_24557 dT47857 dT44368 dT40535 dT37150 dT48611 dT13498 TS7_195845122 dT39089 dT1951 dT46613 St_32307 dT27775 dT9233 dT20879 dT50316 dS_60240 dS_2186 dT21991 dT36711 dT35867 dT38873 dS_60100 dT24515 dS_69530 dT18538 St_18499 dT51420 dT44682 dT52029 dT49010 dS_65237 dT40657 S5_59674699 dC_167176 dT48489 St_80242 dT38881 dT46307 dT48485 dT35147 dS_10139 dT19568 dT51425 S2_226559654 dT29846 dT28837 dT8910 dT29482 dS_20460 dT43417 dS_10652 dT51 S2_230011940 dT15831 dT48530 dT24512 dT16591 115 St_8592 dS_27918 TS3_102980429 S5_52320341 dT3180 dT25619 dT45821 dT12987 dT38040 St_26736 dT37638 dS_68998 dT24372 St_20919 dT3891 dT5081 dC_223673 C3_64621327 dS_64991 dT9987 dT32542 S1_6210777 dT38020 dT22674 dT32245 St_77447 dT11606 St_26737 dT1466 dT20435 TS2_230011891 dT12425 dT50307 St_77432 dT332 dT10713 dT50566 St_32659 TS2_230011940 S4_199859831 dT30922 dT43632 dT23060 120 dT3328 dT47787 St_30275 dS_292 dT12488 dT34571 dS_63173 S3_64306230 dT16346 dT43387 dT44853 dS_16469 dT3675 dS_2101 C1_10467974 dS_68159 St_32493 dT10355 C4_182430864 S7_213184560 dC_45050 dT34797 dT3930 dT24042 St_20865 dT41621 dT22858 dT2954 dT1702 dT6794 St_50788 St_20885 dC_15179 dT32036 dT24014 125 dT3027 dT14931 dT25239 dC_64162 dC_23137 dT29409 dT5877 dC_42488 dT53493 dS_20958 dT25838 dT20060 dT19025 dT41922 dT38795 dT46041 dT15989 dS_43657 St_11615 dT7711 dT14847 dT31457 dS_33008 dT42777 dC_88689 St_46107 dT39297 dT5281 dT9151 dS_58337 dT29788 dT51942 dT16799 dT42995 dT20166 dT51180 dT16992 dS_53679 dT30560 dT7436 dT47137 S3_45974916 130 dT2680 dT38805 dT21277 dT27325 dT5896 dT37923 dT20483 dT44577 dT38911 dT15721 dT48120 St_36519 dT23539 dT36538 dT16233 C3_202051529 dT25304 dT8867 dT13850 St_4579 dS_64711 dT24470 St_22003 St_39105 dT28463 dS_37868 dT6835 dT8152 St_13438 dT29992 dT4710 135 dT18370 dT17665 dT15649 dT19734 dS_55275 dT14945 St_857 C6_39611566 dT38486 TS2_210211525 dS_63047 St_84477 dT35938 dT14717 dT32534 S1_211915643 St_795 dT6810 dS_69156 dT6193 dT3588 dT30050 S6_35473070 dT18085 dT20949 St_46582 140 St_5465 dT38488 S3_258397449 TS2_102571886 dT33996 G8_2178054 S5_31045825 S1_55738760 dT12622 dT5844 dS_38908 dT11037 S3_16736775 dT50361 dT47966 dS_53100 dT6548 S3_41221675 dT5902 dT43430 St_1128 dT41013 dT33250 dT44728 dT47917 St_8964 dT590 S3_260011343 dS_68599 St_70710 St_10547 dT45550 St_58349 C4_54525188 dS_896 dT2555 C2_79892636 St_46842 dT3437 dT159 dC_299059 St_46843 St_44026 dT27467 dT31503 dS_68339 St_43733 dT50779 dT2426 dT42418 dT14768 dT10674 dT31430 dT44125 dT7777 dT10631 St_17805 dT47502 TS6_10754586 dT12632 St_41094 C3_124558897 dT43605 dS_42852 dT38576 dT44700 dT53273 dT53963 dT25565 dT5708 dT46533 dT13489 dT53333 dT17348 dT33550 dT39834 dT11902 St_22721 dT30409 dT47765 dT51639 dT10926 dT11173 S7_61864734 dT29622 dT32441 dT13526 dT640 dT25986 dT13896 dT49885 dT21007 dT19710 dT37138 dT17365 dT37626 dT53592 dT24945 dT36198 dT37696 dT38592 dT2179

Figure 2-10. Consensus genetic linkage map of napiergrass. 81

Figure 2-11. Circos plot of the mapped TASSEL de-novo UNEAK napiergrass markers with pearl millet reference genome. Pearl millet pseudomolecules start with “PM” and are color coded for each pseudomolecule. Napiergrass linkage groups start with “LG” and are in green color. Each line corresponds to tags that showed significant BLAST hits to the pearl millet genome (> 80% identity and > 50 bp length).

Figure 2-12. Syntenic regions between napiergrass linkage groups and the pearl millet genome. PM01 to PM07 are pearl millet pseudomolecules, LG01 to LG14 are napiergrass linkage groups. The small dots represent significant BLAST hits of mapped UNEAK tags to the pearl millet genome (>80% identity and >50 bp length).

CHAPTER 3 MAPPING QTLS CONTROLLING FLOWER NUMBER AND FLOWERING TIME IN NAPIERGRASS

Introduction

Napiergrass (Cenchrus purpureus Schumach) is a tropical perennial grass that originated from Africa (Singh, Singh, and Obeng 2013) and is an important fodder crop widely used as feed for dairy cows (Farrell, Simons, and Hillocks 2002). In addition, due to its high biomass potential, napiergrass is considered as a promising crop for cellulosic biofuel with higher dry biomass yield compared to sorghum (Sorghum bicolor), maize (Zea mays), sugarcane (Saccharum sp.), switchgrass (Panicum virgatum), johnsongrass (Sorghum halepense), and Erianthus (Ra et al. 2012). Napiergrass was introduced to the United States in 1913 (Burton 1990). Being non-edible, napiergrass escapes the 'food versus fuel' debate as the biofuel feedstock and has competitive advantage over tree species because it can be harvested for biomass in the first year after planting. Furthermore, the lignin content, which is considered a hindrance to the fermentation of biomass to ethanol, is much lower in napiergrass than woody biomass,

10% in napiergrass compared to 20-30% in woods (Tong, Smith, and Mccarty 1990;

Mckendry 2002). In addition, napiergrass tolerates multiple harvests after which it shows better ratooning ability than energycane (Cuomo, Blouin, and Beatty 1996). This supports a constant feedstock supply and minimizes transportation costs of biomass.

Napiergrass is a short-day plant and flowering in tropical climates occurs from autumn through winter (Singh, Singh, and Obeng 2013). Early flowering cultivars produce abundant wind dispersed seeds, which contribute to the high potential of invasiveness or weediness (D’Antonio and Vitousek 1992; Loope, Hamann, and Stone

1988; Schofield 1989). Therefore, The Florida Exotic Plant Pest Council has listed napiergrass as an invasive species (FLEPPC 2011). Controlling flowering or modifying flowering time of napiergrass can minimize invasiveness and boost its potential as biofuel feedstock. Late flowering will reduce the total seed production and low temperatures during late season may even compromise development of viable seeds

(Grabowski et al. 2016), thus reducing invasiveness. Significant genotypic variation for agronomic traits have been documented in napiergrass (Sinche et al. 2018). Flowering time showed a considerable variation within an F1 population of napiergrass, thus there is a good reason to believe that improvements of flowering time in napiergrass can be achieved using genetic approaches. Genome editing tools could be utilized to regulate flowering time (Jung et al. 2018; Jung and Müller 2009) and to avoid unintended spreading of napiergrass. However, routine transformation and genome editing protocols still need to be developed for napiergrass (S. Zhou et al. 2018; J. Wang et al.

2017). So far there are no quantitative trait loci (QTL) mapping reports for agronomic traits in napiergrass. Even in this genomic era, there is no noticeable genomic data publicly available for napiergrass. This presents a challenge to improve agronomic traits in napiergrass breeding utilizing marker assisted selection (MAS). Therefore, more genomic resources and a better understanding of the genetic basis of flowering time is necessary in napiergrass.

Most studies on napiergrass were limited to assessing genetic diversity and relatedness using random amplification of polymorphic DNA (RAPD), amplified fragment length polymorphism (AFLP), isozymes, and simple sequence repeats (SSRs) developed for other species like pearl millet (Cenchrus glaucum) and buffelgrass

(Pennisetum ciliare) (Lowe et al. 2003; Bhandari, Sukanya, and Ramesh 2006; Harris-

Shultz, Anderson, and Malik 2010; Kandel et al. 2016; Dowling et al. 2013; Dowling,

Burson, and Jessup 2014; López et al. 2014; Smith et al. 1993). Environmental biosafety of napiergrass can be increased by utilizing molecular markers that are linked to specific traits of interest such as late flowering. This MAS allows effective selection of breeding materials. Identification of QTLs related to flowering will support MAS programs for breeding napiergrass with delayed or less flowering characteristics, limiting invasive potential of napiergrass. This may enhance the potential for utilizing napiergrass as a forage and biofuel crop.

QTL mapping tries to identify stretches of DNA that are closely linked to genes underlying the trait of interest by performing statistical analysis of genome-wide molecular markers and traits measured in progeny of controlled crosses (Stinchcombe and Hoekstra 2008). The advancement of next generation sequencing (NGS) has hugely facilitated QTL identification and mapping. High density genetic maps developed by genotyping-by-sequencing (GBS) have been used successfully to identify genes related to flag leaf traits in wheat (Triticum aestivum) (Hussain et al. 2017), winter hardiness and fall dormancy in Medicago sativa (Adhikari et al. 2018), bunch fruit weight and height in palm (Elaeis guineensis) (Pootakham et al. 2015), and bloom date in peach (Prunus persica) (Bielenberg et al. 2015). QTLs related to flowering time have also been identified in other species as well. In pearl millet, QTLs for flowering time co- mapped with QTLs for stover yield, grain yield, and biomass yield in LG4 and LG6

(Yadav et al. 2003). Similarly, three QTLs on LG2, LG3, and LG4 were identified for grain yield across variable post-flowering moisture environments in pearl millet (Bidinger

et al. 2007). In wheat and barley, vernalization (Vrn) and photoperiod (Ppd) genes were involved in flowering time variations (Cockram et al. 2007). Transcription factors such as

AP2 and agamous-like MADS-box were detected in Adzuki bean for QTLs related to flowering time, maturity, and seed coat color (Y. Li et al. 2017). In rice, several different

QTLs for flowering time were identified and cloned in early and late flowering cultivars that may have been involved in adaptation to cold regions (Izawa 2007; Xue et al.

2008). However, no QTL analyses in napiergrass have been reported so far. Whether the orthologs in other species underlying flowering time also control the flowering time in napiergrass is unknown.

The variation of flowering time in napiergrass accessions can be exploited to understand the genetic basis of flowering in napiergrass. QTL analyses in napiergrass is enabled by the generation of the first genetic map of napiergrass, which contains

1,913 SNP markers called from GBS of a pseudo-F2 mapping population, derived from a cross between an early flowering line and a late flowering line of napiergrass (Paudel et al. 2018). This map facilitates the identification of QTLs for various traits including flowering related traits that can be utilized in the future for MAS. The main objective of this research was to use this first genetic map of napiergrass to identify markers linked to genes controlling flowering time and flower number through QTL analyses.

Materials and Methods

Development of a Mapping Population

A mapping population of 185 F1 hybrids was developed by crossing an early flowering accession N122 and a late flowering accession N190 of napiergrass (Sinche

2013). Accession N190 produced a higher biomass and had a reduced number of thick tillers compared to N122 (Sinche 2013).

The 185 F1 hybrids along with the two parents, N122 and N190, and an established cultivar Merkeron were phenotyped as replicated single row plots at PSREU in 2011 and 2012. Clones of the whole population were also planted at the Everglades

Research and Education Center (EREC) in Belle Glade, FL in 2015 and flowering date was recorded in the following year after planting.

Phenotyping the Mapping Population

The field experimental design followed a randomized complete block design

(RCBD) with three replicates of 187 lines (185 F1 hybrids and 2 parents). One block contained a single plant as a replicate of each line. Within each block the lines were randomly assigned to estimate block effects. The first flowering date and flower numbers of each line were recorded in October ~ December of 2012, 2013 on the plants established at PSREU. The first flowering date was also recorded on the plants established in EREC in October ~ December of 2016. The flowering date, defined as the date when the first flower was visible, was documented weekly from the first week of

October to the first week of December. Flowering traits in 2012 and 2013 at PSREU were obtained from a previous study (Sinche 2013). Flowering time (FT) was calculated as the number of days between the first appearance of the flower and vernal equinox

(March 20) for the specific year (Lambert, Miller-Rushing, and Inouye 2010).

Genetic Map

The genetic map was generated based on the SNPs from GBS of the 185 lines of the population (Paudel et al. 2018). Briefly, GBS was used to genotype the 185 F1

individuals and the two parents. SNPs were identified using various software tools to construct a linkage map for maternal and paternal parent by employing a pseudo testcross strategy (Paudel et al. 2018). For the female parental line, a total of 899 SNP loci mapped on 14 linkage groups with a total length of 1,555.17 cM were used. Similarly, for the male parental line, a total of 1,073 markers that were grouped into 14 linkage groups spanning a length of 1,939.19 cM were used (Paudel et al. 2018).

QTL Analysis

For QTL detection we chose the composite interval mapping method on

WinQTLcart 2.5 (S. Wang et al. 2005). Mean values of each trait across three replicates from different years were used for the QTL analysis. The walking speed chosen for all traits was 2 cM. A forward and backward stepwise regression method with a probability of 0.1 and a window size of 10 cM were utilized to determine cofactors. LOD thresholds used to determine the significance of identified QTLs was identified by using the thousand-permutation test to each data set (p ≤ 0.05) (Churchill and Doerge 1994).

Adjacent QTLs on the same chromosome for the same trait were considered different when the support intervals did not overlap (Haggard, Johnson, and St. Clair 2015). The

95% confidence interval was calculated for each QTL considering a 2-LOD support interval (van Ooijen 1992). The QTL span was delimited using LOD-1 support interval

(LSI). The contribution rate (R2) was calculated as the percentage of phenotypic variance explained by each QTL in proportion to the total phenotypic variance. QTLs were named according to McCouch et al. (McCouch et al. 1997). Specifically, the QTLs detected for number of flowers on the linkage map constructed for early flowering parent

N122 were designated “qNFE” (qtl Number of Flowers Early) followed by a linkage

group number and QTL number on the same linkage group for the same trait, separated by a dash “-“. Similarly, the QTLs detected for flowering number based on the linkage map constructed for late flowering parent N190 were named “qNFL”. The QTL for flowering time on linkage map for early flowering parent N122 were named as “qFTE”

(qtl Flowering Time Early) and the QTLs for flowering time on map for late flowering parent N190 were given “qFTL” (qtl Flowering Time Late). QTLs with a positive or negative additive effect for a trait imply that the increased or decreased phenotypic value is contributed by the QTL, respectively.

We categorized the identified QTLs as either stable and potential QTLs. QTLs detected based on the phenotypic data from more than one year were considered as stable, while those detected based on only one year’s phenotypic data were considered as potential QTLs. The positions of the QTLs identified on each linkage map were indicated by using MapChart 2.3 (Voorrips 2002).

Candidate Gene Identification

Sequences were extracted from the sequence tags, which generated the SNP markers flanking QTL region. The extracted sequences were then BLASTed against the pearl millet genome v1.1 (Varshney et al. 2017) to identify nucleotide matches of these sequences to identify the QTL sequence intervals. Gene models and KEGG annotation

(Varshney et al. 2017) of pearl millet genes within the identified QTL sequence interval were extracted. The function of each gene model was checked for its role in flowering and a list of potential candidates was curated.

Results

Phenotypes

Number of flowers

The number of flowers of the 185 F1 individuals, and their parental lines, early flowering line (N122) and a late flowering line (N190) ranged from 0 to 46 with an average of 6 in 2012 (Figure 3-1, Table 3-1) and from 5 to 256 with an average of 61 in

2013 (Figure 3-2, Table 3-2). Many F1 individuals actually didn’t flower before the harvest date (Dec. 6th, immediately before the first predicted frost) in 2012.The correlation between number of flowers for 2012 and 2013 was 0.62 (Fig 3-3). The heritability estimates for number of flowers ranged from 79% to 85% (Table 3-1).

Flowering Time

The days to flowering of the whole F1 population including the two parental lines, ranged from 219 to 270 days with an average of 240 days in 2012 (Fig 3-4) and from

220 to 258 days with an average of 238 days in 2013 (Table 3-1, Fig 3-5). The average number of days to flowering in 2016 was 233 days (Fig 3-6, Table 3-1).

The correlation between days to flowering for 2012 and 2013 was 0.34, while that between 2012 and 2016 was 0.22. Correlation of days to flowering between 2013 and

2016 was 0.26 (Fig 3-7). The heritability for the days to flowering were estimated ranging from 60% to 87% (Table 3-1). Days to flowering was negatively correlated with number of flowers (r = - 0.42 for 2012 and r = -0.64 for 2013, Fig 3-8 and Fig 3-9).

QTL Analysis

Number of flowers

Three stable QTLs (qNFE-1-1, qNFE-1-2, and qNFE-1-3) were identified on LG1 of the genetic map for early parent N122 for number of flowers (Table 3-2, Figure 3-10). 91

Three potential QTLs (qNFE-2-1, qNFE-6-1, and qNFE-6-2) were identified on LGs, 2 and 6 (Table 3-2). The most important number of flowers QTL (qNFE-1-2) for N122 parent (R2=0.20, LOD=10.76) was found in LG1and was located at 123.7 ~ 126.8 cM.

The same linkage group harbored the two additional stable QTLs (qNFE-1-1 and qNFE-

1-3) that explained 15-17% of variance with LOD score ranging from 8.13 to 9.10 at peak intervals from 133.5 cM - 134.6 cM and 116.9 cM - 118.2 cM. This suggests that

LG1 plays an important role for number of flowers in napiergrass.

Five potential QTLs for number of flowers were identified on LG 1, 4, and 5 of the N190 map (Fig 3-11, Table 3-2). These potential QTLs had a PVE of 6-10% with a

LOD value ranging from 3.18 - 4.78. Four out of those five QTLs had additive effects in favor of trait value. One potential QTL (qNFL-11-2) had a negative effect (Table 3-2).

Flowering time

A total of six potential QTLs were detected for FT. Two potential QTLs out of the six were detected on the data of year 2013 on LG 1 of N190 and two potential QTLs were detected on the same LG of N122 with R2 ranging from 0.11- 0.14 and LOD score of 5.40-6.77. Two additional potential QTLs were identified on LG 7 of map N122. The

R2 for these potential QTLs ranged from 0.09 - 0.12 with LOD scores ranging from 3.91-

5.57. Both QTLs showed negative effect of small value (R2 = 0.07) on the trait.

Candidate Genes

Candidate genes were only identified in the genome regions corresponding to the three stable flower number QTLs: qNFE-1-1, qNFE-1-2, and qNFE-1-3. Based on

BLAST analysis of sequences flanking the QTL marker against the pearl millet genome

(Varshney et al. 2017), three genome intervals spanning a genome size of 4 Mb, 12.37

Mb and 1.4 Mb, respectively, were identified on pseudomolecule 1 of the pearl millet genome, corresponding to the three QTLs. A total of 624 gene models were identified within the three QTL intervals of the genome (160 gene models on qNFE-1-1, 397 on qNFE-1-2, and 67 on qNFE-1-3). 295 of these models had KEGG annotation results (71 on qNFE-1-1, 191 on qNFE-1-2, and 33 on qNFE-1-3) (Varshney et al. 2017). This region harbored potential candidate flowering genes such as AGAMOUS, DELLA, Floral homeotic protein DEFICIENS, PPM1 (MIKCC MADS-domain protein), WRKY, and

SERK1 (Table 3-4).

Discussion

Napiergrass is an important forage crop with high potential as a cellulosic biofuel feedstock. Breeding of non- or late flowering varieties that have an extended vegetative growth is not only important for obtaining a high yield but also for reducing invasiveness.

So far, genes involved in flowering regulation have not been identified in napiergrass, which has severely hindered the improvement of flowering time traits in napiergrass breeding. Over the last few years, mapping of QTLs for economically important traits and genome assembly has largely facilitated the breeding programs by the development of MAS (Y. Wang et al. 2011). Recently, a high-density genetic map of napiergrass has been developed (Paudel et al. 2018) that will ease identifying QTLs.

Exploring QTLs is important because many studies have identified candidate loci near

QTLs. For example in cereals, a QTL controlling heading date variation was close to loci controlling photoperiod and vernalization (Laurie et al. 1994; Bezant et al. 1996). In rice, four genes controlling flowering time (Hd1, Hd6, Hd3a, SE5) were identified from 14

QTLs (Yano et al. 2001). In maize, it was shown that flowering time variation was the

result of cumulative effect of several small QTLs and by using 5,000 RILs, a total of 36

QTLs were identified for days to anthesis (Dell’Acqua et al. 2015). Flowering time QTLs in brassicas mapped to similar regions in homologous chromosomes within and between species (Lagercrantz et al. 1996; Osborn et al. 1997). Flowering related genes in Brassica showed a close sequence similarity to one end of chromosome 5 of

Arabidopsis, which contained many flowering related genes (Bohuon et al. 1998). This type of syntenic results showed that the same QTLs may exist in other related species.

The availability of the reference genome of pearl millet (Varshney et al. 2017) largely facilitates the comparative genomic approaches for napiergrass as these two species share one genome (Jauhar and Hanna 1998).

In this study, we phenotyped a biparental F1 or pseudo-F2 population that segregated for flowering time to identify QTLs related to flowering. The traits studied showed a continuous variation indicating that the traits are quantitative in nature. We observed a negative phenotypic correlation between the number of flowers and flowering time in the F1 population. Days to flowering and number of flowers were consistent in different years and locations for the population. The heritability for flowering date ranged from 0.60-0.87 for napiergrass. It was higher than the heritability reported for heading date in rice which ranged from 0.37 to 0.55 (L. Zhou et al. 2016) and lower than that reported for maize (0.82 to 0.93) (C. Wang et al. 2010). For linkage analysis, we employed a pseudo-testcross strategy. Linkage analysis of quantitative traits in outcrossing polyploid species by using single dose markers (1:1) is a common practice due to the limitation of the linkage analysis software (K. K. Wu et al. 1992). In this approach, a pseudo-testcross strategy using heterozygous markers for one parent

and recessively homozygous markers in other parent is employed for linkage map construction (S. Wu et al. 2010; W. Zhou et al. 2015). This strategy has been successfully applied in tree plants such as Pinus elliottii and P. caribaea (Shepherd et al. 2003), legumes such as alfalfa (Adhikari et al. 2018), as well as grasses such as orchard grass (W. Xie et al. 2011), and sugarcane (Yang, Islam, et al. 2018).

In this study, three stable QTLs for number of flowers were identified on the linkage map constructed for N122. An initial study done in pearl millet identified QTLs related to grain number and panicle number in LG1 of pearl millet (Bidinger et al. 2007;

Yadav et al. 2003). However, the lack of shared markers did not allow a definitive conclusion if these two QTLs represent the same locus.

Several QTL regions related to flowering time and number of flowers were located at the end of the LG 1 of napiergrass. Across all years, the early flowering QTLs with the greatest effect was found on LG 1 that correspond to pseudomolecule 1 of pearl millet (Varshney et al. 2017). This highlighted the importance of this chromosome for flowering and it might harbor genes that control flowering in napiergrass. QTL identification has enabled us to link variations at the trait level to variations at the sequence level. Since a QTL can harbor tens to hundreds of genes (Gelli et al. 2016), the identification of genes responsible for phenotypic variation poses a major challenge.

Accurate identification of the underlying genes responsible for trait variation in QTL regions is difficult in napiergrass due to the lack of genomic and transcriptomic resources in napiergrass. Nevertheless, comparative genomic approaches utilizing reference genome of pearl millet has helped us to identify putative candidate genes related to flowering in napiergrass. Several candidate genes affecting flowering time in

model plants have been discussed previously (Kuittinen, Sillanpa, and Savolainen 1997;

Bouché et al. 2016). In rice, QTL analysis of flowering time identified genes such as

Hd1 (CO orthologue in Arabidopsis), Hd6 (CK2 orthologue in Arabidopsis), Hd3a (FT orthologues in Arabidopsis), and SE5 (HY1 orthologue in Arabidopsis) (Yano et al.

2001).

In our study, we located important flowering time related genes including

AGAMOUS, DELLA, DEFICIENS, PPM1 (MIKCC MADS-domain protein), WRKY, and

SERK1 (Table 3-4) in the QTL area of napiergrass that showed sequence similarity to the pearl millet genome. The AGAMOUS gene encodes a transcription factor that regulates genes to determine stamen and carpel development (Yanofsky et al. 1990).

DELLA protein plays a key role in negative regulation of gibberellin biosynthesis that regulates many cellular and developmental events include flowering, pollen maturation, and the transition from vegetative growth to flowering (Yoshida et al. 2014). DEFICIENS is an ortholog of APETALA3 in Arabidopsis that functions in petal and stamen organ identity (Zahn et al. 2005). PPM1 is ubiquitously expressed throughout vegetative and reproduction tissues and may have diverse functions (Singer, Krogan, and Ashton

2007). WRKY has been related to several abiotic responses and accelerates flowering by regulating FLOWERING LOCUS T and LEAFY (Phukan, Jeena, and Shukla 2016).

SERK1 (SOMATIC EMBRYOGENESIS RECEPTOR-LIKE KINASE1) acts as a negative regulator of abscission metabolism and is required for anther development

(Lewis et al. 2010). Further investigations of these genes involved in flowering in wild accessions of napiergrass are necessary to elucidate the basis for intraspecific variation which is of great relevance to breeders. Future fine-mapping analysis by identifying

additional genetic variants in the QTL regions or by targeted resequencing of the QTL intervals in a subset of individuals can lead towards accurate identification of targets for improving flowering time in napiergrass and related crops. To verify the napiergrass

QTLs detected in our analysis, phenotyping should be repeated in multiple locations.

Breeders are more interested in QTLs that have large effect (Kearsey and Farquhar

1998) and the QTLs identified in this study can be used in the future for MAS in order to breed late flowering napiergrass.

Conclusion

In summary, this study reports for the first time QTL analysis in napiergrass.

Three stable QTLs controlling number of flowers in napiergrass were identified, which can explain 15%-20% of the phenotypic variation. We also identified three potential

QTLs controlling flowering time in napiergrass, which can explain 11%-14% of the phenotypic variation. Gene models in pearl millet that mapped to two stable QTLs harbored MADS-box transcription factors such as AGAMOUS and DEFICIENS, along with other proteins such as DELLA, WRKY, and SERK1 that are involved in flowering time regulation in plants. The QTLs detected in this study will be valuable information for napiergrass breeding programs and help to understand the genetic basis of flowering.

This study confirms that flowering time and flower number are highly heritable traits in napiergrass. Therefore, the late flowering napiergrass genotypes developed by Sinche

(2013) constitute a valuable germplasm resource to develop late flowering napiergrass cultivars. In addition, validation of the putative candidate genes identified in this study should lead to targets for genome editing to manipulate flowering time in napiergrass.

Figure 3-1. Histogram of number of flowers in the mapping population in 2012 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

Figure 3-2. Histogram of number of flowers in mapping population for 2013 in Citra, FL. X-axis represents the number of flowers and y-axis represents the count of plants. Average number of flowers of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

200

100 Number of Flowers 2013 of Number

0 0 10 20 30 40 Number of Flowers 2012

Figure 3-3. Scatterplot of number of flowers between 2012 and 2013 in Citra, FL. X- axis represents the number of flowers in 2012 and y-axis represents the number of flowers in 2013. Data is adapted from Sinche, 2013.

100

Figure 3-4. Histogram of number of days to first flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

101

Figure 3-5. Histogram of number of days to first flowering in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted. Data is adapted from Sinche, 2013.

102

Figure 3-6. Histogram of first flowering date in the mapping population in 2016 in EREC, Belle Glade, FL. X-axis represents the number of days to flowering (FT) and y-axis represents the number of accessions. Days to first flowering of two parental lines, N122 and N190 are plotted.

103

2 R 0.11 250

240

230

Days to Flowering Daysto 2013 220 220 230 240 250 260 270 Days to Flowering 2012

R2 0.04 250

240

230

220 Days to Flowering Daysto 2016

220 230 240 250 260 270 Days to Flowering 2012

R 2 0.06 250

240

230

220 Days to Flowering Daysto 2016

220 230 240 250 Days to Flowering 2013

Figure 3-7. Scatterplot of first date of flowering between different years and locations. X-axis and Y-axis represent the number of days to flowering (FT) in different years. Data for 2012 and 2013 are from Citra, FL and adapted from Sinche, 2013. Data for 2016 are from Belle Glade, FL.

104

20 Number of Flowers 2012 of Number 10

220 230 240 250 260 270 Days to Flowering 2012

Figure 3-8. Scatterplot between number of flowers and days to flowering in 2012 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2012 and y-axis represents the number of flowers per line in 2012. Data is adapted from Sinche, 2013.

105

2 R 0.41 200

100 Number of Flowers 2013 of Number

220 230 240 250 Days to Flowering 2013

Figure 3-9. Scatterplot between days to flowering and number of flowers in 2013 in Citra, FL. X-axis represents the number of days to flowering (FT) in 2013 and y-axis represents the number of flowers per line in 2013. Data is adapted from Sinche, 2013.

106

Linkage map of male parent N122

LG1 LG2 LG6 LG7

S3_66915283 0.0 dT11902 0.0 S4_192354592 0.0 S2_174467255 0.0 S3_66915258 0.6 dT17348 0.5 qNFE-2-1 dT6838 5.5 dT46533 1.3 dT43261 3.2 dT44555 6.5 dT53963 2.5 dT27134 6.9 dT38576 3.2 dT3062 5.4 S3_66915278 7.5 C3_124558897 3.3 S1_7068544 8.9 dT10674 4.2 dT20572 7.7 dT3437 10.2 TS6_10754586 4.7 dT9434 9.3 dT46011 10.4 dT10631 4.8 dT38476 13.3 dT27467 6.0 dT35347 13.8 dT2555 8.6 dT33533 14.1 dT45550 9.1 dT50501 14.2 S1_203215691 14.4 dT28618 14.7 dT11037 10.5 C2_174467255 dC_58121 dT39822 15.3 S5_31045825 10.8 dT13383 16.9 dC_58157 18.9 dT28250 15.4 S6_35473070 12.0 dT1325 20.5 C5_27253496 16.0 dT35938 14.3 dT24339 21.7 TS7_74798545 16.5 C6_39611566 14.8 dT25619 20.8 dT33835 22.9 dT31263 19.7 dT6835 15.9 dT15831 23.7 dT10459 23.8 dT6267 21.1 dT16233 20.7 dT29846 26.5 dT21208 24.3 dT30050 21.8 dT38911 21.0 dT20879 dT3030 dT41386 24.5 27.1 C1_119093525 C1_119093531 22.0 dT38805 23.3 dT27514 dT16455 26.4 dT37316 22.3 dT6794 24.4 dT40535 dT18538 27.3 dT38126 26.9 dT12054 22.6 dT14931 25.2 dT48489 27.7 dT20154 27.2 S1_71014695 24.0 dT16992 25.7 C4_200698121 28.2 dT24498 28.0 dT17414 24.9 dT46041 26.6 dT4430 28.9 qFTE-7-1 dT29057 25.8 dT16346 29.2 dT23119 29.7 S3_284635389 26.2 dT47787 29.6 dT43805 30.0 dT2475 27.9 dT42989 32.8 dT3891 30.9 dT18370 dT12390 28.0 dT50566 34.0 dC_13461 31.7 dT50610 37.5 dT41113 28.1 dT9987 36.3 S4_116088663 33.3 dT30582 28.3 dT13255 37.0 dC_618193 30.4 dT32416 38.2 dT44347 41.9 dT44577 31.3 dT29726 41.8 C4_90405411 39.4 dT6436 42.3 dT34797 32.9 dT8320 39.9 dT4353 43.7 dT51180 33.5 dT23975 44.8 S4_29656682 41.2 dT28611 36.3 dT31396 46.2 dT42579 44.4 dT35867 37.1 dT43235 47.3 dT4620 44.5 dT23729 40.2 dT5027 48.8 dT2439 49.2 dT1631 47.5 dT34784 41.0 dT23432 49.4 dT37028 41.5 dT39900 51.7 S1_193475897 41.8 dT39167 52.1 dT43233 52.7 dT861 52.5 dT36388 42.3 dT18870 53.3 dT23905 53.4 dT3762 42.5 dT8454 53.8 dT42142 55.0 dT7751 55.4 S7_227510 55.3 dT38724 48.8 dT35142 56.8 TS1_74139971 dT11106 49.2 S7_161179045 S7_161179072 57.7 dT6132 49.9 dT7984 50.2 dT5958 60.6 dT21588 60.1 dT13772 50.6 dT8821 61.9 dT53762 51.2 dT29230 66.0 dT8946 51.8 dT14775 64.4 dT18605 52.3 dT4496 66.5 dT29128 53.6 dT25121 69.1 dT6466 67.8 dT34305 66.6 dT52531 56.0 dT39882 69.2 dT16151 68.9 dT41949 60.2 dT30721 dT19708 69.3 TS4_49707946 69.2 dT5751 69.4 dT47752 68.8 dT36233 69.5 dT11273 70.1 dT18641 71.2 dT43759 70.7 dT28669 71.7 S4_109511769 71.8 dT15498 dC_1941 TS4_79794887 76.3 dC_5486367 dC_428936 74.3 dT36845 76.6 dT50064 dT41460 dT33420 77.6 dT5222 74.5 dT25340 77.7 dT17613 77.8 dT43295 74.8 dC_124707 79.3 dT31359 75.0 dT31777 84.2

dT44933 77.3 S4_17921468 85.4 qNFE-6-1 dT33834 81.5 dT39406 85.9 dT7877 83.4 dT27185 84.6 dT46138 82.2 dT45232 86.6 dT4301 C4_13275000 87.4 dT24376 89.5

qNFE-6-2 dT7992 88.3 dT36267 89.7 dT22362 88.7 dT26600 90.2 S7_168784407 90.3 dT987 91.3 dT52456 91.1 S4_21013305 91.5 S7_168784355 91.7 dT17187 92.4 dT10193 91.7 dT34107 92.0 dT3299 95.9 dT22232 92.6 dT29166 96.5 dT9256 94.2 S4_17796331 95.7 dT2229 98.5 dT26372 98.5

dT19151 108.9 dT22571 110.8 dT8342 114.3 S7_146917990 109.0 dT52469 116.0 dT36501 111.1 dT37356 117.2 dT11509 111.6 dT35128 118.2 dT4155 111.9 dT1086 112.6 dT53136 118.6 S7_100737594 114.2 dT39139 120.7 qNFE-1-1 dT20499 112.8 dT36338 122.1

qFTE-1-1 dT19693 124.8 S6_118469241 119.5 dT46088 118.4 dT44366 125.5 dT14954 121.1 dT36827 120.7 dT28974 126.0 dT39428 123.1 dT36253 121.3 S1_249594415 126.5 qNFE-1-2 dT19912 124.3 dT5535 127.0 dT53574 124.7

dT8792 127.2 qFTE-1-2 dT43321 127.6 C7_38012075 S7_52134705 124.9 dT42596 127.4 dT34936 127.9 dT48305 S7_52134761 126.6 128.0 dT2297 129.9 dT32706 129.2 dT8340 131.0 S7_52134764 128.5 dT49414 129.9 qNFE-1-3 dC_234325 132.0 C1_251517594 130.2 dT20274 134.9 dC_259973 130.3 dT22921 135.8 S1_251483789 130.4 dT30440 139.1 dT41404 130.8 dT14807 140.0 dT25902 136.4 dT42403 131.3 dT30703 141.4 dT43823 137.3 dT9863 131.6 dT6973 142.1 S1_260924412 133.5 dC_196800 142.9 dT35453 134.2 dT18761 143.2 dT45297 134.7 dT36412 143.8 dT28492 135.8 dT24998 144.5 dT47753 136.7 dT25533 145.0 dT19286 145.8 dT32523 137.3 dT29143 145.7 dT43692 146.0 dT21677 137.7 S6_55827481 146.2 dC_2450640 138.4 S5_1337832 147.8 dT48962 139.3 dT23441 149.3 dT50248 139.6 dT28159 150.8 dT27136 140.5 dT9688 151.5 dT32605 141.2 dT43184 153.6 C6_200342258 154.5 S1_267892930 149.1 dT11814 dT9665 155.1 156.0 dT16180 dT18516 156.7 157.3 dT26922 dT18539 157.3 157.7 S7_119895303 dT13751 158.5 159.3 dT8614 dT50350 160.4 160.3 dT29758 161.1 dT22426 163.1 dT10112 162.1

dT17747 168.9

dT45130 171.6 dT29238 173.2 dT6553 173.8

qFTE-7-2

dT13151 177.6 dT37830 178.0 S7_66630588 179.5 dT39337 180.7 dT44668 183.3 dT25362 183.6 dT31763 184.3 dT28445 183.8 dT9471 185.2 dT3618 186.5 Figure 3-10. Linkage map of male parent N122 showing potential and stable QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group. 107

Linkage map of male parent N122

LG1 LG4 LG11

dT20387 0.0 S1_6210777 0.0 dT49930 0.0 dT41575 3.7 dT1702 dT3027 4.6 qFTL-1-1 dT909 4.8 S4_131830044 G4_131830067 6.6 S4_131830067 6.7 C1_10467974 5.2 qNFL-1-1 dT38795 5.4 S5_128073888 4.6 S4_131830066 7.2 S3_64306230 5.8 S4_131830063 7.3 S1_6210761 8.3 qNFL-4-1 dT3328 6.1 dT24515 11.1 dT22876 8.3 dT49252 6.4 dT17924 dT12269 9.4 qFTL-1-2 15.6 dT30456 6.5 TS5_114098156 16.0 C4_131830077 9.5 dT16886 6.7 dT16773 16.2 dC_61912 dC_55022 9.7 dT53891 7.0 dT40995 16.5 dT22567 10.3 dT16764 7.2 dT15817 16.8 dT26116 10.7 dT35147 8.3 dT14651 18.3 dT38883 11.0 dT49010 8.6 dT18480 dT36340 dT10040 11.3 18.8 qNFL-4-2 dT2473 8.9 TS5_128053793 dT13827 dT26884 11.6 dT1951 9.6 C5_156998414 C5_112498054 19.1 dT10739 dT621 11.7 dT16527 9.9 dC_24267 dT17325 S7_89250999 11.9 C1_220011116 12.0 dT27345 dT30788 19.2 dT17578 12.0 dT50264 20.4 dT9315 19.4 dT25711 12.1 dT29470 21.6 S5_40839446 19.6 dT5186 12.3 dT32431 22.3 dT4737 20.2 dT8749 12.6 dT9029 22.5 dT31217 22.1 dC_10270 13.2 dT50826 22.7 dT32027 22.2 dT649 13.7 dT47548 23.0 dT46427 22.4 dT53247 13.9 dT45924 dT7985 dT32695 23.0 dT53022 dT12194 14.0 dT24786 dT732 23.5 S7_39646339 15.6 C1_102656250 23.6 S7_39646309 15.9 dT13338 23.7 S7_39646292 16.2 dT33677 16.7 dT19473 23.8 dT43636 29.3 dT18485 23.9 dT3510 17.3 dT46948 24.3 dT19298 17.9 dT51201 24.5 dT1762 21.7 dT41196 22.3 dT23838 24.8 dT31735 33.3 dT23364 25.2 dT30018 22.7 dT23237 30.6 dT40278 23.3 S7_83881022 23.7 dT28498 38.0 dT7491 25.5 dT51779 38.5 dT50145 38.1 dT38871 28.1 dT43458 39.3 dT29364 35.0 dT6993 41.3 dT29597 42.0 dT43048 41.0 dT5196 43.9 C5_132196063 41.2 dT14180 44.3 dT20788 41.7 dT48800 44.6 dT32207 41.9 dT3431 45.2 dT24792 43.5 dT48497 45.3 S1_17257463 45.8 dT46652 46.4

dT9167 46.1 qNFL-11-1 dT28049 46.3 dT48243 46.6 dT27717 49.6 dT15391 50.1 S1_17257559 46.9 dT8910 50.9 S1_17257551 47.2 dT1524 51.2 dT18424 47.8 dT14375 53.4 dT973 48.2 dT31410 54.0 dT8459 50.7 dT31722 54.3 dT33654 50.8 dT48112 51.1 dT53672 52.0 dT12272 53.7 dT18219 59.0 dC_223673 59.4 qNFL-11-2 dC_359278 58.8 dT36949 59.7 C1_133709193 58.9 dC_339765 59.0 S1_171369178 59.2 dT43632 63.3 dT40041 59.7 dT24276 63.6 dT34571 64.0 dT18064 61.2 dT3675 64.6 dT971 61.9 S7_213184560 64.9 dT22725 62.2 dT22858 65.4 dT10843 dT19448 62.3 dT32036 66.2 dT47057 63.4 dT29409 67.8 C1_234266014 65.4 TS1_273230575 66.0 dT30844 66.4 dT19025 71.6 dT7314 67.2 dT7856 72.9 dT7711 72.5 dT766 68.5 dT39297 74.1 dT36979 69.2 S1_217117736 73.5 dT5489 dT30297 75.1 76.2 dT43892 dT3813 78.0 77.6 dT42995 78.6 dT47137 80.2

dT48559 82.4

S1_241274562 87.0

dT37923 90.6 dT42534 91.8

dT49410 93.8 dT52149 94.9

dT32206 99.6 dT6664 101.1

dT8114 104.1 dT53440 105.2 dT40421 108.2 dT14806 108.5 dT22647 108.9 S5_31904873 109.7 dT39799 110.2 dT3053 110.6 S5_26740824 111.4 dT4924 111.7 dT37762 113.4

dT26721 118.4 dT26387 119.1 dT6733 119.8 dT17794 120.1 S5_16898738 120.9 dT37016 121.4 dT47995 123.0 dT34310 123.7 dT51527 124.4 TS5_13509703 124.6 dT54119 125.1 dT53202 125.4 dT32125 126.4 dT27794 126.6 dT2435 127.9 dT46426 129.8 dT20384 130.1 C5_3132406 131.6 dT15117 131.9 dT29259 132.8 dT780 133.3 dT15952 134.3 dT26168 135.2 dT36121 137.1 dT32758 139.0 dT30470 142.4 Figure 3-11. Linkage map of female parent N190 showing potential QTLs. QTLs for number of flower (NF) are color coded in blue and that for flowering time (FT) are colored in green. A black bar means a GBS-SNP marker. Markers are labeled on the left and genetic distance (centimorgan as unit) are labeled on right. LG = Linkage group.

108

Table 3-1. Descriptive statistics of flowering date and number of flowers for 185 F1 hybrids of a cross (N190  N122) at Citra, FL., in 2012, 2013 (Sinche 2013) and 2016.

Statistic Flowering datea Number of flowers (plot-1) b

2012 2013 2016 2012 2013 Min 10/25 10/26 10/23 0 5 Max 12/06c 12/03 12/02 46 256 Mean 11/15 (240) d 11/13 (238) d 11/07 (232) d 6 61 SE 1 d 0.5 d 0.05 d 0.76 2.86 H2 0.60 0.87 0.64 0.85 0.79 a Flowering date: mm/dd b 1.8 m long plots. c The harvest concluded before the first predicted frost on 12/06/2012 and several genotypes did not flower until that day. d Value in parenthesis indicates number of days to first flowering counted from March 20 of that year. SE Standard error H2 Broad sense-heritability estimated on entry mean basis

109

Table 3-2. Detailed information about the QTLs for number of flowers. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM. Parent QTL code LG Year Peak markers Peak LOD PVE Allele LSI (cM) Flanking Markers (R2) dir.

116.9 cM - 118.2 N122 qNFE-1-1 1 2012, 2013 dT37356 8.13 0.15 + cM dT52469, dT35128

123.7 cM - 126.8 N122 qNFE-1-2 1 2012, 2013 S1_249594415 10.76 0.20 + cM dT36338, dT5535

133.5 cM - 134.6 S1_260924412, N122 qNFE-1-3 1 2012, 2013 dT35453 9.10 0.17 - cM dT45297

N122 qNFE-2-1 2 2013 TS6_10754586 3.48 0.08 + 3.1 cM - 4.8 cM dT38576, dT10631

N122 qNFE-6-1 6 2012 S4_17921468 3.64 0.06 + 84.3 cM - 86 cM dT31777, dT45232

N122 qNFE-6-2 6 2012 dT10193 3.45 0.06 + 91.1 cM - 91.9 cM dT52456, dT34107

N190 qNFL-1-1 1 2013 dT3328 4.04 0.10 + 5.6 cM - 6.7 cM dT38795, dT53891

S5_128073888, N190 qNFL-4-1 4 2012 S1_6210761 3.66 0.07 + 7.6 cM - 8.8 cM dT24515

N190 qNFL-4-2 4 2012 dT40995 4.78 0.10 + 16.2 cM - 16.7 cM dT16773, dT15817

N190 qNFL-11-1 11 2013 dT8910 4.22 0.08 + 47.5 cM - 52.9 cM dT29364, dC_223673

N190 qNFL-11-2 11 2013 dT43632 3.18 0.06 - 57.9 cM - 64.2 cM dT8910, dT3675

110

Table 3-3. Detailed information about the QTLs for flowering time. LG = Napiergrass linkage groups, Peak markers represent the marker at the highest peak of QTL, LSI represents the LOD-1 support interval in cM. Parent QTL Linkage Year Peak Peak PVE (R2) Allele dir. LSI (cM) Flanking Markers code Group markers LOD

118.3 cM - 120.8 N122 qFTE-1-1 1 2013 dT53136 6.77 0.14 + cM dT35128, dT39139

126.5 cM - 128.3 S1_249594415, N122 qFTE-1-2 1 2013 dT48305 5.40 0.11 - cM dT32706

N122 qFTE-7-1 7 2012 dT44347 3.91 0.09 - 31 cM - 42.1 cM dT24498, dT6436

178.3 cM - 182.1 N122 qFTE-7-2 7 2016 dT39337 5.57 0.12 - cM dT37830, dT44668

N190 qFTL-1-1 1 2013 dT41575 3.60 0.07 - 3.3 cM - 4.6 cM dT20387, dT3027

dT2473, N190 qFTL-1-2 1 2013 dT1951 3.33 0.07 - 8.9 cM - 11.2 cM C1_220011116

111

Table 3-4. List of putative flowering related genes from the genome of pearl millet (Varshney et al. 2017). Gene name Description Gene ID Chr Start position End position Exon QTL count AGAMOUS AGAMOUS-like protein; Pgl_GLEAN_10032931 chr1 268933236 268935744 7 qNFE-1-3 K09264 MADS-box transcription factor DELLA DELLA domain GRAS Pgl_GLEAN_10026530 chr1 248097573 248099492 1 qNFE-1-2 family transcription factor rga-like protein; K14494 DELLA protein DEFICIENS Floral homeotic protein Pgl_GLEAN_10000416 chr1 250352485 250350575 6 qNFE-1-2 DEFICIENS, putative; K09264 MADS-box transcription factor PPM1 PPM1; MIKCC MADS- Pgl_GLEAN_10029410 chr1 253208367 253209680 1 qNFE-1-2 domain protein PPM1; K09264 MADS-box transcription factor WRKY K13424 WRKY Pgl_GLEAN_10027907 chr1 242557234 242558164 3 qNFE-1-2 transcription factor 33 SERK1 SOMATIC Pgl_GLEAN_10000999 chr1 250195809 250194274 2 qNFE-1-2 EMBRYOGENESIS RECEPTOR-LIKE KINASE 1; kinase/ transmembrane receptor protein serine/threonine kinase

112

CHAPTER 4 EVALUATE THE GENETIC BACKGROUND OF FLOWERING TIME IN A NAPIERGRASS GERMPLASM COLLECTION

Introduction

Napiergrass (Cenchrus purpureus) is an important forage and biofuel candidate

(Singh, Singh, and Obeng 2013). However, there has been relatively little effort in improving napiergrass as compared to cereal crops for increased yield and agronomic characteristics (Wanjala et al. 2013). Napiergrass cultivation in the US is limited by its potential for invasiveness (Sollenberger et al. 2014). Typically, napiergrass is very persistent and only few diseases like napiergrass stunt disease and napiergrass head smut disease have been reported with significant impact on biomass yield or crop persistence (Farrell, Simons, and Hillocks 2002).

Existing genetic diversity of napiergrass needs to be assessed in order to strengthen breeding programs for development of high yielding and late flowering cultivars that can be considered as non-invasive (Wanjala et al. 2013; Sinche 2013).To initiate this, there is a need to have accurate information on the available germplasm to identify clones that form the base population for breeding late flowering lines. The first step of breeding for flowering time is to create variability within the germplasm so as to widen the base population of the candidate clones for selection (Juma 2014). Variability can be due to intrinsic reasons or due to random genetic drift, natural selection, mutation, gene flow and transfer (Y. Xu et al. 2017). This variability can be determined by assessing the level of polymorphism between accessions and within each accession.

This information can be used for clone development. To determine the variability of the available germplasm, stable and informative molecular markers are critical to study the

113

genetic diversity. Molecular characterization of germplasm will increase breeding efficiency by providing important genetic information of the breeding materials and will ultimately help in identification and introgression of desirable agronomic traits into high yielding clones through marker assisted selection (MAS). This will lead to increased napiergrass yield with decreased invasiveness.

Napiergrass cultivar discrimination has been mostly done based on morphological and agronomic characteristics that lead to inconsistency in identification of specific accessions (Struwig 2007). Consequently, many cultivars of napiergrass are described with more than one name/identifier. Marker informed elimination of redundancy in collections of napiergrass accessions will facilitate the maintenance of the collection and its use in crop improvement programs. In this regards, molecular markers have proven to be very effective to distinguish between morphologically related individuals of the same species (Wanjala et al. 2013).

Hence, genetic assessment of napiergrass accessions with molecular markers was initiated several decades ago. Characterization of different napiergrass germplasms has been carried out using Restricted Fragment Length Polymorphism

(RFLP) (Smith et al. 1993), Random Amplification of Polymorphic DNA (RAPD) (Lowe et al. 2003), Amplified Fragment Length Polymorphism (AFLP) (Harris-Shultz,

Anderson, and Malik 2010; Wanjala et al. 2013), inter-simple sequence repeats (ISSRs)

(de Lima et al. 2011), sequence-related amplified polymorphism (SRAP) (X.-M. Xie et al. 2009), and simple sequence repeat (SSR) markers (Azevedo et al. 2012; López et al. 2018; López et al. 2014). Currently, single nucleotide polymorphism (SNP) markers are gaining popularity as they are ubiquitous in the genome, occur in a large number

114

with high density, relatively inexpensive, codominant, and can act as functional markers if located in coding regions (Rafalski 2002; Ganal, Altmann, and Röder 2009). Further, genome wide SNP markers can be used as diagnostic tools for fingerprinting germplasm, to correct discrepancies that are inherent in taxonomic methods based on morphological characteristics and to estimate genome composition of clones (Yang,

Song, et al. 2018).

Flowering time is very important to plant breeders because agronomic traits like biomass yield and quality depend on the timing and intensity of flowering (Veeckman et al. 2016). Flowering time is a key factor in plant adaptation and is linked to various attributes like plant height, yield, and number of leaves (Durand et al. 2012). In many species, flowering is induced in response to day length. Different flowering responses are categorized as short-day, long-day, intermediate-day, or day-neutral based on the day length requirement (Schlegel 2009; Bastow and Dean 2002). Napiergrass belongs to the short-day plants (Osgood, Hanna, and Tew 1997; Singh, Singh, and Obeng

2013), in which flowering is favored by day lengths shorter than the corresponding nights.

Early flowering in perennials limits vegetative growth, and delayed flowering will hinder seed development to progress completely before the onset of adverse weather

(Grabowski et al. 2016). Delaying flowering could potentially lead to increases in biomass yields as seen in other grasses like switchgrass where a 10 days delayed flowering increased biomass by > 25% (Price and Casler 2014). Flowering time is an important target trait in napiergrass breeding programs because different flowering time helps to adapt cultivars in different geographic locations or seasons. It is also important

115

because flowering has been associated with invasiveness of napiergrass in Florida

(FLEPPC 2011). High number of wind-dispersed seeds has contributed to categorizing napiergrass as an invasive species in Florida, thus limiting its utilization as an industrial crop for forage or biofuel (López et al. 2014). For this reason, development of cultivars that are late flowering is critical for napiergrass management in the field because late flowering lines can be harvested before flowering, as development of viable seeds is impaired by low temperatures in Florida during Nov-Feb. Studying flowering time variation in napiergrass can help to identify potential genes that promote or hinder seed formation in napiergrass. Mining genome sequences that are available for several grass species for flowering related genes and their characterization can also identify candidate genes. These candidate genes can be investigated in different populations for late flowering. In addition to this, integrating genomics with conventional breeding will shorten the breeding cycle for selection.

The substantial variations in flowering time within the population of napiergrass can be exploited to not only select lines with desirable traits but also to identify sequence variations that are associated with it or could be causal. By characterizing the relationships between genetic variations and flowering time we might be able to identify loci affecting flowering time variation in napiergrass. Flowering time can be controlled by a large number of quantitative trait loci (QTLs) with small effects (Buckler et al. 2009).

Genetic and genomic approaches that have been used in other species to characterize flowering time genes might not be feasible in napiergrass due to its large genome size and high polyploid level. This challenge can be mitigated by using genome reduction approaches like exome-capture sequencing methods (Evans et al., 2014). Exome

116

sequencing targets only genic regions of the genome with high depth and can produce gene-level resolution of genome-wide patterns of sequence variation (Grabowski et al.

2016).

To acquire a global perspective of sequence variations and to obtain an accurate population structure of napiergrass germplasm, we deeply sequenced the coding regions of 94 accessions from the Tifton nursery. Then we used genome-wide SNPs identified to evaluate the genetic relationships of the germplasm as well as performed genome wide association studies targeting flowering time in napiergrass.

Materials and Methods

Plant Materials and Phenotyping

A total of 94 napiergrass accessions from the germplasm collection available at the USDA-ARS, Tifton nursery were used in this study. Nodes from those accessions of napiergrass were cut and planted in 6” pots on August 19, 2015 in Everglades

Research and Education Center (EREC), Belle Glade, FL. Transplanting of the pots was done on October 12, 2015. Plots were established in a randomized complete block design with three replications and spacing of 4 ft x 4 ft plant to plant as well as row to row distance were maintained. Emergence of the first flower was noted every week during October – December 2016. Days to flowering was calculated as the number of days between the first appearance of the flower and vernal equinox (March 20)

(Lambert, Miller-Rushing, and Inouye 2010).

DNA Extraction

Young and healthy leaf tissues were harvested from each individual of the germplasm collection. DNA extraction was done following the protocol described

117

previously (Dellaporta, Wood, and Hicks 1983). The extracted DNA samples were run on a 2% agarose gel to check the quality and quantified with PicoGreen to meet the requirements of exome sequencing.

Targeted Candidate Genes

Genes related to flowering in different monocots and other species were mined using various publications (Table 5-1). In addition, Refseq database was searched for

“flowering” and filtered for “Green plants” to get potential flowering related genes (Pruitt,

Tatusova, and Maglott 2007). All gene models present in the pearl millet genome

(Varshney et al. 2017) were used for probe design to achieve a genome-wide coverage of all coding sequences (CDS).

Probe Design

The protein sequences for flowering genes were clustered using CD-HIT

(Weizhong Li and Godzik 2006). Probes of 120 bp long with an overlap of 60 bp were developed using Emboss 6.5.7 (Rice, Longden, and Bleasby 2000). These probes were mapped back to the pearl millet genome (Varshney et al. 2017) using BLAT (Kent

2002). A hit was defined under cutoff: e-value ≤ 1e-05; alignment identity = alignment length * percentage of identity ≥96 (120 bp * 80%=96 bp). Probes that hit the genome less than three times were selected and probes containing repeats were removed. 2kb promoter region and 500 bp downstream region of the flowering genes were also included for probe design with the same criteria.

A relatively relaxed criteria was chosen to target missing flowering time related genes. For this, 120 bp long probes were designed with 90 bp overlap and good probes were selected with the criteria (Alignment * Identity) / 120 >= 80. The probes that hit the

118

genome less than 4 times were selected and probes containing repeated regions were removed.

Probes were also designed to target the coding regions of the whole genome.

For this, sequences of 38,579 genes of pearl millet were used to retrieve 120 bp probe sequence with maximum allowable overlap of 30 bp. Good probes were selected under the criteria alignment * identity / 120 >= 80. Probes with multiple hits on the genome or containing repeat regions were removed. A maximum of 2 probes were kept for each

CDS.

A relaxed criterion was implemented for genes, where no probes were designed according to the standards above. For those genes, probes were designed with 90 bp overlap and good probes were selected as alignment * identity /120 >= 80. Probes with up to 2 hits on the genome were kept and probes containing repetitive regions were discarded.

Probe Synthesis, Selection, and Sequencing

The designed probes were submitted to RapidGenomics LLC (FL, USA) for quality check. Probes that passed quality criteria of RapidGenomics were then selected using the following criteria: for flowering related genes, a total of 18 probes were targeted per gene while for genome wide CDS, 1 probe per gene was targeted. Four probes were targeted per gene for 2 kb promoter region and 1 probe per gene was selected for 500 bp downstream region of the gene. Finally, a total of 37,000 probes were submitted for synthesis. The synthesized probes were used to capture the DNA fragments of the germplasm collection. The captured DNA fragments were sequenced using the Illumina HiSeq 3000 platform (150 bp paired-end reads). The probe synthesis,

119

library preparation, target enrichment, and sequencing were performed by

RapidGenomics LLC (FL, USA).

Sequence Read Trimming and Mapping

Raw reads were trimmed using Trimmomatic (Bolger, Lohse, and Usadel 2014).

Reads that contained more than 50% low-quality bases (Q20) were removed and adaptor sequences were trimmed. Trimmed reads were aligned to the pearl millet genome v1.1 (Varshney et al. 2017) using BWA-mem (H. Li and Durbin 2010). For SNP calling, we retained only uniquely mapped reads with a combination of mapping quality

‘0’ and ‘XA:Z’ tag (both indicates multiple alignment).

SNP Calling

SNPs were called using GATK 3.2.2 (McKenna et al. 2010) with settings (-T

UnifiedGenotyper -glm BOTH -R -ploidy 4 -mbq 20). For each sample, homozygous

SNPs required a depth of ≥11; heterozygous SNPs required a depth of ≥31; and at least

2 reads for minor allele were required to make a call. At the population level a mapping quality of ≥30; minor allele frequency ≥0.05; and call rate of ≥95% was employed.

Population Structure

In order to infer population structure of the germplasm accessions, we used

DAPC available in the adegent package for R (Jombart 2008). Groups were clustered using k-means clustering, where the best k minimized the Bayesian Information

Criterion (BIC). GWASpoly (Rosyara et al. 2016) was used with six different models for

GWAS using sequence variants that had a call rate ≥ 95% and MAF ≥ 0.05. Molecular

Evolutionary Genetics Analysis version 6.0 (Tamura et al. 2013) was used to infer genetic relationships among the germplasm collection and to perform phylogenetic

120

analysis. Less than 5% alignment gaps, missing data, and ambiguous bases were allowed at any position. Neighbor-Joining method was used to infer the evolutionary history. The evolutionary distances were computed using the p-distance method and are in the units of the number of base differences per site (Tamura et al. 2013).

Results

Candidate Genes Related to Flowering

A total of 3,542 flowering time related genes were curated from different organisms including Arabidopsis, Brachypodium, maize, rice, Setaria, sorghum and switchgrass (Table 4-1). This list included 1,285 flowering related genes retrieved from the Refseq database (Pruitt, Tatusova, and Maglott 2007). CD-HIT clustered the flowering genes into 1,279 non-redundant genes that were further compared to peptides from the pearl millet genomes using BLAT. In pearl millet, 506 genes had BLAT hits to the non-redundant flowering amino acid sequences. Thus the 506 genes were targeted for the probe design.

Probe Design

The CDS sequences of 506 ortholog genes in pearl millet related to flowering were subjected to probe design. A total of 8,314 probes were designed and 4,908 probes targeting 473 pearl millet flowering time related genes remained after filtering.

The number of probes designed per gene ranged from 1-91 (Figure 4-2) and the number of probes designed per gene was proportional to the size of the gene (Figure 4-

3). To further cover the genes without probe designed, the selection criteria were relaxed (see Materials and Methods) and an additional 9,099 probes were designed targeting 422 additional genes.

121

We extracted 2kb region upstream of the flowering genes to design probes for the promoter regions. A total of 3,215 probes were designed targeting promoter regions of 243 pearl millet genes. We further extracted 500bp downstream of the flowering genes and designed probes from the terminator regions. A total of 1,025 probes were designed from the terminator regions targeting 239 flowering related genes.

Probes were also designed to target the coding regions of the whole genome.

Thus the sequences of 38,579 gene models annotated in the pearl millet genome v1.1

(Varshney et al. 2017) were included to design genome wide probes. The length of the genes ranged from 100 bp to 5600 bp (Figure 4-1), with most of them in the range of

500 bp -1600 bp. A total of 53,525 probes targeting 29,015 out of 38,579 genes were designed. For the genes that couldn’t be targeted, we relaxed selection criteria and further design 7,306 probes that targeted additional 4,318 genes (Materials and

Methods).

Thus, a total of 78,988 probes were totally designed and were submitted for quality check, after which 63,967 probes passed the quality test provided by Rapid

Genomics. From these probes, a total of 37,000 unique probes were further selected for synthesis and sequencing. This final probe set targeted a total of 818 flowering related genes with a maximum of 18 probes per gene. These genes were distributed on all chromosomes of the pearl millet genome (Figure 4-4). The probe set also targeted a total of 28,519 genome wide CDS with 1 probe per CDS. Promoter region of 240 genes were targeted with a maximum of 4 probes per gene and terminator region of 231 genes were targeted with a maximum of 1 probe per gene.

122

Sequence processing

A total of 375,135,615 paired-end reads (2×150 bp) were generated by Illumina sequencing. The number of paired-end reads per sample ranged from 2.3 M to 5.6 M, with an average of 3,990,804 read pairs per sample (Figure 4-5). A total of 354,016,510

(94.35%) read pairs passed the quality control. From the good quality reads, a total of

20,901,032 raw SNPs were called by GATK (McKenna et al. 2010). After filtering for quality, depth, minor allele frequency (MAF), and call rate, a total of 78,129 high quality

SNPs remained that were subjected to genome wide association analysis using

GWASpoly (Rosyara et al. 2016).

A total of 2,893 SNPs (one SNP per 10 Kbp) were selected to perform DAPC.

Based on the Bayesian Information Criteria (BIC), the germplasm could be clustered into 3 clusters (k=3) (Figure 4-6). Using these SNPs, the germplasm could be projected into three linear discriminants (LD)s (Figure 4-7) that corresponded to their geographic origin. Group 1 had accessions originating from Africa, group 2 had accessions that originated from America (central and south America), and group 3 had a mix of lines from Asia and America. However, three accessions whose origin were in Taiwan (Asia) were grouped into the same group as accessions originating from Africa (group 1).

Seven accessions whose origin was listed as America were also grouped into the

African group 1. Group 2 exclusively contained accessions originating in America.

Group 3 contained accessions from the Americas and Asia, with one exception that came from Africa.

123

Phylogenetic analysis

The phylogenetic analysis involved 94 samples in MEGA6 by using 2,872 SNPs

(one SNP every 10 Kbp). Positions that had less than 95% site coverage were removed. Evolutionary analyses showed that most of the lines could be assigned based on the geographic locations, however, there were some mixtures (Figure 4-8). Group 1 mostly contained lines that originated in Africa, while group 2 contained lines with origin from America. Many accessions in group 2 had unknown origin. Group 3 included a mix of African, Asian, and American lines. Asian lines were grouped in group 1 and group 3.

Genome wide association analysis

Average days to flowering in the germplasm collection ranged from 232 days to

279 days (Figure 4-9). The heritability for days to flowering was 0.58. All the high-quality

SNPs after filtering (78,129) were used for genome wide association analysis by using

GWASpoly with six different models (Figure 4-10, Figure 4-11). For every model, the majority of p-values lie close to the 1:1 dashed line (Figure 4-10). For days to flowering

(DTF), the strongest association signal was observed with the 1-dom-alt model (Fig 4-

11) for a marker on Chromosome 5 at location 129,365,754 with an estimated marker effect of -16.79. Other models didn’t give significant associations (Fig 4-11).

Subsequently, we explored the region 100kb upwards and 100kb downwards of this marker to identify potential candidate genes. Ten gene models were located within this region and we identified potential flowering candidate genes including calcium-binding protein CML and enoyl-CoA hydratase in this region.

124

Discussion

Dissecting genetic loci controlling flowering time is critical for napiergrass breeding as flowering is related to invasiveness and biomass accumulation. In this research, we applied TES for high throughput genotyping of tetraploid napiergrass. We were able to investigate sequence variants using TES which reduced genome complexity by selectively sequencing only the coding regions. After applying quality filter, we identified 78,129 high quality SNPs in the germplasm collection that serve as valuable and novel genomic resource for napiergrass breeding programs. Moreover, we performed GWAS in the napiergrass germplasm collection using these SNPs and identified one SNP as well as two candidate genes that was associated with flowering time.

We applied DAPC to analyze the population structure of the napiergrass germplasm collection. DAPC is a multivariate method to analyze genetic structure and does not rely on Hardy-Weinberg equilibrium assumptions or linkage disequilibrium assumptions (Jombart 2008). We selected 2k high quality and evenly distributed SNPs for the analysis and inferred that three groups existed in the germplasm collection. This is in contrast with results from (Kandel et al. 2016) who inferred 5 groups in a subset of this population. Those results were based on only 29 SSR markers in contrast with our study that used 2k uniformly distributed SNPs. In this study, we had 24 lines that were not used in the previous study (Kandel et al. 2016). Out of these, 14 lines were of unknown origin. Our study grouped accessions N43 and N51 into the same group similar to previous study that grouped biofuel type and naturalized napiergrass into

125

different groups (López et al. 2014). However, our results grouped N13 in a separate group compared to N43 and N51.

We identified a total of 20 million raw SNPs and compiled a comprehensive sequence variation database for napiergrass that included 78k high quality SNPs. This dataset provides accurate genotypes for the germplasm collection and serves as the foundation for population genomics analyses in napiergrass, like gene mapping and

GWAS. We were able to identify potential candidate variations using GWASpoly

(Rosyara et al. 2016), which is primarily designed for autopolyploid species, however, it has also been successfully used in diploids, as well as allopolyploids like wheat (Phan et al. 2018).

The regulation of flowering time is important for plant adaptation and biomass production. In this study, we used genotypes derived from exome-capture sequencing and flowering time information from a germplasm collection of napiergrass to characterize the genetic architecture underlying flowering time regulation in napiergrass.

This is the first GWAS study performed in napiergrass. Recently, QTL mapping was conducted in a bi-parental mapping population of 185 progenies with multiple QTLs detected for flowering time and number of flowers (Chapter 3). We detected one significant QTL in this study on chromosome 5 of the pearl millet genome, which corresponded to linkage group 4 of napiergrass, where two QTLs (qNFL-4-1 and qNFL-

4-2) for number of flowers were identified in the previous study (Chapter 3). Other QTLs identified in Chapter 3 were not identified by GWAS. This could be explained by the type of population studied (germplasm collection vs. bi-parental population), statistical method employed (GWAS vs. linkage mapping) and strict threshold to declare

126

significance (p<0.05 after Bonferroni correction vs. logarithm of odd <3) (Yang, Islam, et al. 2018). In the 200kb region surrounding the most significant marker, we identified two genes related to flowering. Calcium-binding protein CML cause alterations in flowering time and can act as a switch in response to day length perception and the gain of CML function results in early flowering (Tsai et al. 2007). Similarly, enoyl-CoA hydratase is important for seed germination (Richmond and Bleecker 1999). Other potential flowering candidates like phytochrome B and AP2-like factors were located at 150 kb and 165 kb respectively away from the significant marker. The markers and genes associated with flowering time could be fully utilized in napiergrass breeding programs after they have been validated in a much larger population.

Identifying genes that are responsible for trait variation is a major challenge (Gelli et al. 2016). Due to the unavailability of transcriptome sequences for napiergrass, it will be difficult to pinpoint the underlying genes responsible for trait variation. Future evaluations could lead us towards potential targets for improving flowering time in napiergrass and related crops. Currently there is no report that has utilized TES in napiergrass and we have made this first attempt to use exome sequencing for calling variations in napiergrass for flowering related genes as well as performed GWAS.

GWAS in polyploids is challenging because most studies are based on diploid genetics and software that are modeled using diploid genetics (Bourke et al. 2018). While there is no software available specifically to perform GWAS in allotetraploids, recently, two software were released that accept polyploid data, namely GWASpoly (Rosyara et al.

2016) and SHEsisPlus (Shen et al. 2016). In this study, we found 1 significant SNP using GWASpoly. Modeling GWAS studies based on polyploidy as compared to diploidy

127

can result in identification of more associated SNPs (Ferrão et al. 2018). Therefore, the selection of a proper model is critical in GWAS. In the future a software that is particularly modeled for allopolyploidy might become available and may help to identify more SNPs associated with the trait. Moreover, we were also limited by the size of the population. Our population of 94 accessions might not have enough power to detect the variations in flowering time, which is mostly controlled by minor effect QTLs (Dell’Acqua et al. 2015). In addition, phenotyping may lack accuracy since it was only completed for one year and at one location. These factors might have played a role in the limited number of SNPs identified for the flowering time. Phenotyping the germplasm collection at multiple locations and for multiple years may allow to identify significant associations to the traits of interest. Further, genotyping more accessions of napiergrass germplasm could also help in identifying other important variations that were not detected in this study. In addition, this germplasm collection can also be phenotyped for other traits like those related to biomass accumulation to associate these traits with sequence variations and potential candidate genes.

Probe design is critical for improving the performance of TES for deep sequencing. Probes can be designed from transcribed sequences (expressed sequence tags or RNA-seq reads) or genomic sequences (whole genome sequencing, reduced representation library, genotyping by sequencing) of the species of interest or related species. We used the genome of a closely related species of napiergrass, pearl millet

(Varshney et al. 2017), to design probes for this study. Using 37k probes, we targeted

818 flowering genes, and 28,519 overall genes according to pearl millet genome.

128

Since napiergrass is allotetraploid with two similar sub-genomes, most of the probes designed for this study might have multiple hits. With the unavailability of reference genome of napiergrass, we are unable to separate the probes that could have come from each sub-genome. While TES provides high coverage data, some non- primary targets could also be captured, which is unavoidable for organisms that do not have a reference genome (Peng et al. 2017).

Conclusion

Our study showed the feasibility to applying targeted enrichment sequencing with probes designed from CDS of pearl millet genome to target the genomic regions in napiergrass. We inferred the structure of the germplasm collection as well as constructed phylogeny for the germplasm. TES allowed us to perform a GWAS on this population, and we identified one significant SNP and two candidate genes for flowering time variation. The success of this approach could be improved by increasing the number of accessions studied, phenotyping for multiple years at multiple locations, as well as using the reference genome of napiergrass once it becomes available. With the availability of the genome reference, we can assess uniqueness of probes that will increase unique mapping rate as well as minimize the proportion of off-target capture.

129

Table 4-1. Genes and publications related to flowering used in this research. Organism Genes Reference

Arabidopsis 306 (Bouché, Lobet, Tocquin, & Périlleux, 2016)

Brachypodium 23 (Higgins, Bailey, & Laurie, 2010)

Maize 692 (Li et al., 2016)

Maize 148 (Xu et al., 2012)

Rice 201 Wiki pathways

Rice 39 (Lee & An, 2015)

Setaria 53 (Mauro-Herrera et al., 2013)

Sorghum 141 (Mace, Hunt, & Jordan, 2013)

Switchgrass 654 (Grabowski et al., 2016) mRNA 1,285 Refseq db: flowering + greenplants

Grand Total 3,542

130

80 70 60 50 40 30

Number ofgenesNumber 20 10

900-999 100-199 300-399 500-599 700-799

1900-1999 2900-2999 3900-3999 4900-4999 1100-1199 1300-1399 1500-1599 1700-1799 2100-2199 2300-2399 2500-2599 2700-2799 3100-3199 3300-3399 3500-3599 3700-3799 4100-4199 4300-4399 4500-4599 4700-4799 5100-5199 5300-5399 Length of gene (bp), n=1213

Figure 4-1. Histogram of length of flowering related genes.

131

Numberofgenes 20

0 0 20 40 60 80 100 Number of probes designed per gene

Figure 4-2. Number of probes designed per gene.

132

Figure 4-3. Number of probes designed as a factor of the size of the gene.

133

Figure 4-4. Distribution of the targeted flowering genes in the genome of pearl millet. Each blue bar represents a flowering gene mapped on the pearl millet genome.

134

6,000,000

5,000,000

4,000,000

3,000,000

2,000,000 Number ofpairs readNumber 1,000,000

0 Sample

Figure 4-5. Number of paired-end reads per sample.

135

Value of BIC

versus number of clusters

450

440

BIC

430 420

5 10 15 20

Number of clusters

Figure 4-6. Bayesian Information Criteria (BIC) vs. number of clusters in k-means clustering suggests K=3 in the germplasm collection.

136

0 LD2

-2

-4

-2.5 0.0 2.5 5.0 LD1

Figure 4-7. Projection of the napiergrass germplasm collection using first two linear discriminants (LDs) from discriminant analysis of principal components (DAPC). The shape of the points represent grouping by DAPC (circle = group 1; triangle = group 2; square = group 3) and the colors represent the origin continent: black = Americas; blue = Asia; and cyan = Africa.

137

NILRI-14984 G1

NILRI-16786 G1

N190 G1 N157 G1

N109 G1 N199 G2 N228G2 N51 G2 N210 G2 N43 G2 N43 N222 G2 N203 G2 N13 G1 N-Mer G2 N212 G2 N70 G1 N198 G2

N8 G2 N7 G1 N114 G2 N22 G1 N223 G2 N19 G1 N214 G2 N238 G1 N215 G2 N226 G1 N116 G2 N188 G1 N204 G2 N127 G2 N155 G1 N211 G2 N72 G1 N225 G2 N71 G1 N-6X G2 N161 G1 N205 G2 N180 G1 N128 G2 N129 G2 N151 G1 N12 G2 N166 G1 N132 G2 N150 G1 N74 G3 N138 G1 N130 G3 N137 G1 N9 G3 N178 G1 N14 G3 N185 G1 N37 G3 N239 G3 N182 G1 N122 G3 N186 G1 N168 G3 N170 G1 N16 G3 N183 G1 N20 G3 N172NILRI-16791 G3 G1 N181 G1

N164 G1 N40 G3 N171 G1 N35 G3 N36 G3 N163 G1 N34 G3 N152 G1 N42 G3 N39 G3

N131 G1 N41 G3 N179 G1 N32 G3 N23 G1 N147 G1 N24 G1 N66 G1

N56 G1

N69 G1 N69 N67 G1 N67 N240 G1 G1 N68

N242 G1

N244 G1

N243 G1

Figure 4-8. Evolutionary relationships of taxa. G stands for group assigned by DAPC and are in different shapes (circle = group 1; triangle = group 2; square = group 3) and the fill colors represent the origin: black = Americas, blue = Asia, cyan = Africa, and white=unknown).

138

230 240 250 260 270 280 Days to flowering

Figure 4-9. Histogram for days to flowering trait in napiergrass germplasm collection.

139

Figure 4-10. QQ plots for different models using GWASpoly using 78,129 SNP markers. DTF = Days to flowering.

140

Figure 4-11. Manhattan plots for different models using GWASpoly. P values adjusted with FDR at 0.05. DTF = Days to flowering.

141

CHAPTER 5 GENERATION OF INTERSPECIFIC HYBRIDS BETWEEN PEARL MILLET AND NAPIERGRASS AND EVALUATION OF THEIR PERFORMANCE

Introduction

Napiergrass or elephantgrass (Cenchrus purpureus Schumach) is an important forage crop that is widely used in Africa for dairy cows because of its high yields and nutrient value (Singh, Singh, and Obeng 2013). Mature napiergrass reach plant heights of 5-6 m and up to 20 nodes per stem (Boonman 1988) and out-yield other grasses by a significant margin (Ra et al. 2012). Napiergrass is generally propagated via stem cuttings. Clonal propagation increases propagation costs as cultivation is typically done manually. Therefore, even cultivars of napiergrass with high biomass quality like Mott, which produced average cattle gains of 0.97 kgd-1 compared to industry standard of

0.39 kgd-1 of ‘Pensacola’ bahiagrass (Paspalum notatum Flugge) (Sollenberger et al.

1989), have not been used commercially in the US because of the limitations associated with the vegetative propagation (Diz and Schank 1993). Production of large seeds that can be harvested and planted mechanically will propel the commercial success of napiergrass.

Napiergrass is the fastest growing plant in the world (Karlsson and Vasil 1986) with reported dry biomass yield of 45 dry t ha−1 year−1 in Florida (Woodard and

Sollenberger 2012) and as high as 80 dry t ha−1 in tropical countries (Vicente-Chandler,

Silva, and Figarella 1959). Therefore, napiergrass has a great potential as lignocellulosic feedstock for biofuel production. Being non-edible and able to grow on marginal land, napiergrass escapes the 'food versus fuel' debate and has competitive advantage over tree species as it can be harvested for biomass in the first year after

142

planting. Additionally, lignin content, which is considered a hindrance to the fermentation of biomass to ethanol, is much lower in napiergrass than woody biomass,

10% in napiergrass compared to 20-30% in woods (Tong, Smith, and Mccarty 1990;

Mckendry 2002). Most perennial biomass crops like switchgrass (Panicum virgatum L.) have establishment issues due to small seed size, slow growth rate, dormancy, and negative response to high planting densities (Noble Research Institute 2007). Leading bioenergy candidate grasses like switchgrass, Miscanthus, and energycane are not capable of both direct seeding and high production of biomass in the establishment year and in contrast to napiergrass do not tolerate multiple harvests per year (Singh, Singh, and Obeng 2013).

Napiergrass is a short-day plant and flowering in tropical climates occurs from autumn through winter (Singh, Singh, and Obeng 2013). Early flowering cultivars produce an abundant amount of small and wind dispersed seeds making seed collection difficult and increasing its invasive potential (D’Antonio and Vitousek 1992;

Loope, Hamann, and Stone 1988; Schofield 1989). Therefore, napiergrass is listed as an invasive species by the Florida Exotic Plant Pest Council (FLEPPC 2011).

Invasiveness in napiergrass can be effectively controlled by developing interspecific triploid hybrids between napiergrass and pearl millet (Cenchrus americanus, 2n=2x=14)

(Hanna 1981). The chromosomes in the A’ genome of napiergrass are homologous to the A genome of pearl millet (Jauhar 1981) and these two species hybridize naturally to produce pearl millet napiergrass (PMN) hybrids that are triploids (AA’ B genome) and thus are sterile (Singh, Singh, and Obeng 2013). PMN hybrids do not set seed and so, do not pose a threat of uncontrolled establishment through dissemination of seeds

143

(Hanna and Monson 1980) and are not considered invasive (Jessup 2013). PMN hybrids combine superior forage quality of pearl millet and high yielding ability of napiergrass (Gupta and Mhere 1997; Osgood, Hanna, and Tew 1997). Some of these hybrids produced higher biomass (18.9 Mg ha-1) than napiergrass (17.5 Mg ha-1) and pearl millet (13.2 Mg ha-1) in Louisiana (Cuomo, Blouin, and Beatty 1996). Pearl millet seeds are larger in comparison to napiergrass and no seed shattering occurs on the pearl millet panicle (Fig 5-1). The shape and size of PMN seeds are similar to the female pearl millet parent. These seeds can be planted using seed drills. Development of seeded varieties that are sterile will represent a significant step in napiergrass breeding because establishment of fields by seeds will allow automation of planting, thus a significant cost reduction (Osgood, Hanna, and Tew 1997). Normally, the resulting hybrid (2n=21) with AA’B genome has greater similarity to the napiergrass type due to larger genetic contribution (66.7% chromosomes) and dominance of the napiergrass B genome over the pearl millet A genome for genetic characters such as earliness, inflorescence and leaf characteristics (Obok, Ova, and Iwo 2012; Gonzalez and Hanna 1984). Most of the characteristics like resistance to pests, vigorous growth, and high forage yield potential are on the B genome (Hanna 1987).

Both napiergrass and pearl millet are protogynous in nature, a phenomenon where stigmas are exerted prior to anther exertion, therefore, they are predominately cross-pollinated that results in high heterozygosity (Dowling, Burson, and Jessup 2014).

The heterozygous out-crossing nature of napiergrass and pearl millet leads to significant segregation and lack of uniformity in progenies. Because of the heterozygosity in napiergrass and pearl millet, PMN hybrids exhibit a high level of

144

heterosis (Dowling, Burson, and Jessup 2014). For successful commercial application, it is important to get a high level of heterosis in the hybrids while maintaining uniformity in biomass and persistence of the progenies. Uniformity, the ability of a cultivar to produce a specific phenotype instead of a varying phenotype, increases the amount of predictability on the total biomass yield (Makumburage and Stapleton 2011). Using

PMN hybrids for commercial cultivation requires seeds that produce plants with a high biomass yield and a certain level of uniformity. Selection for stand uniformity was found to be associated with increased tolerance to environmental stress in maize (Tollenaar and Wu 1999). PMN hybrid combinations can produce non-germinating seeds

(Aken’ova and Chheda 1973), or produce hybrids with varying yield potential (Hanna and Monson 1980). Phenotypic variation can be due to genetic variation caused by alleles segregating within a population, epistasis and mutations, or environment, caused by fluctuating external condition (Fraser and Schadt 2010).

Commercial production of PMN hybrids can be facilitated by utilizing cytoplasmic male sterile (cms) lines of pearl millet. ‘Tift 23A’, a cms line of pearl millet paved the way to produce seed-propagated PMN hybrids that can facilitate seed harvest (J. B. Powell and Burton 1966). However, the cms lines of pearl millet are dwarf type forages. It is critical to choose the right parental combination in order to maximize yield of interspecific PMN hybrids. For this, the cms trait in the dwarf forage type pearl millet needs to be introgressed into high biomass pearl millet lines and homogenized so that progenies from these will be uniform. The backcross method has been commonly used to transfer entire sets of chromosome from foreign cytoplasm in order to create cytoplasmic male-sterile genotypes (Acquaah 2012). These lines can then be used to

145

cross with inbred napiergrass to produce progenies of male and female sterile PMN hybrids.

Since PMN hybrids can produce forage until the season’s first frost, they have the potential to address the fall forage deficit in the southeast USA (Cuomo, Blouin, and

Beatty 1996). PMN hybrids can be easily established using seeds and may produce high yields in the first year itself. PMN hybrids are more resistant to most pests and diseases than pearl millet (Hanna and Monson 1980). Expressed sequence tags – simple short repeat (EST-SSR) markers can be used in order to confirm triploid hybrids in a breeding program (Dowling et al. 2013). Therefore, developing seeded PMN hybrids with high biomass yield and a certain level of uniformity should have a major impact in the forage and biofuel industry. As such, the impact of different levels of homozygosity/selfing of pearl millet and napiergrass parents on biomass yield and uniformity will inform us about the best strategy to manage field breeding of PMN hybrids. Therefore, it is necessary to identify the level of heterozygosity that is present in the progenies from different crosses by evaluating biomass yield and uniformity of the hybrids under field conditions. There have been no previous studies done on the impact of different levels of selfing of pearl millet and napiergrass parents on the agronomic performance of the PMN hybrids.

Introgression of cms into high biomass pearl millet should facilitate field production of PMN seeds by open pollination of cms pearl millet with napiergrass. In this study we report the biomass yield and uniformity of progenies produced from crosses between four different parental types of female pearl millet (cms forage, high biomass,

146

cms high biomass, and homozygous high biomass) and a napiergrass male parent (25-

17).

Materials and Methods

Description of Male Sterile Line of Pearl Millet

In this research we used a forage type pearl millet, Tift 85D2A4 (Tift 85) which is a cytoplasmic male sterile line that flowers in about 75 days. The cms line of pearl millet

Tift 85 was received from Dr. Wayne Hanna, University of Georgia. The A4 cytoplasm was transferred from a wild subspecies of pearl millet (Wayne Hanna, personal communication). Tift 85D2B1 is used to maintain sterility of Tift 85D2A4 and it is self- fertile with good seed set. It was developed by selecting a rust and leaf spot resistant plant from a selfed population of a BC5 plant developed by backcrossing Tift23D2B1 to a wild grassy introduction [Pennisetum glaucum (L.) R. Br. Subsp. Monodii (Maire)

Brunken] (Hanna, Wells, and Burton 1987).

Production of cms Lines of Pearl Millet

Cytoplasmic male sterility present in Tift 85 was introgressed into three high biomass yielding pearl millet lines based on vigor: PI 288787 01 SD (787), PI 215603 01

SD (603), and DLSBF. PI 288787 01 SD is a late flowering accession collected in India with an average of 19 nodes, 0.880 gm seed weight and 370 cm plant height (USDA,

ARS, and NGRP 1963). PI 215603 01 SD is a late flowering accession collected in India with an average of 370 cm plant height and 0.720 gm seed weight (USDA, ARS, and

NGRP 1954). DLSBF is an African introduction from Burkina Faso (Wayne Hanna, personal communication). Pollen from the pearl millet elite lines was dusted onto stigma of dwarf type cms Tift 85. Thus produced F1 progeny was back crossed to a selfed high

147

biomass line of pearl millet. Male sterility on individual plants were tested by visual observation prior to crossing in each generation. Selfing of parental line and back crossing with the most advanced line of the pearl millet continued. Finally, we were able to introgress cms into one of the three elite pearl millet lines (line 787). We continued to back cross the cms line 787 four times in order to get a cms version of the elite line of pearl millet.

Production of PMN Hybrids

In order to obtain near inbred lines of napiergrass, two lines of napiergrass were selfed, namely, Schank and 25-17. These two lines were selfed two times and flowering was induced by controlling photoperiod in a growth chamber. However, selfed second generation of napiergrass didn’t flower due to problems with photoperiod control in the greenhouse. Therefore, crosses of parental napiergrass (25-17) with different types of female pearl millet lines were performed (Table 5-1). The interspecific triploid (3x=21) hybrids were produced in the growth chamber by pollinating pearl millet female plants with napiergrass pollen.

These four crosses represent the cross of male napiergrass parent (25-17) with four different types of female pearl-millet lines (cms parent – Tift85 [A], selfed 5th generation -787S5 [B], BC4 generation – MS 787 BC4 [C], and original parent – P787

[D]).

Seedlings from these four crosses were grown and evaluated for triploidy based on flowering. Nodes from the confirmed triploids were cut and replanted as replicated clones into 6” pots. Five clones (nodes) for each plant were grown in the greenhouse in

UF/IFAS Plant Science Research and Education Unit (PSREU), Citra, FL.

148

The pots were filled with sandy soil from the field in PSREU to mimic the soil conditions of the field. The nodes were planted on Jan 30, 2018 and were irrigated twice daily. Fertilizer at the rate of 20-20-20 was supplied through irrigation using Dosatron.

After the nodes germinated, the ‘N’ fertilizer rate was increased to 46-0-0 on Feb 21,

2018 in order to promote vegetative growth of the plants. The plants were transplanted to the field on March 28, 2018. The plants were irrigated with 12mm weekly in the absence of rainfall after planting. On April 20, 2018 the plants were fertilized with 33.62 kg ha-1 N; 11.21 kg ha-1 and 33.62 kg ha-1 K plus micronutrient package followed by irrigation. The plants were further fertilized on May 10, 2018 with 84.06 kg ha-1 N; 20.17 kg ha-1 P and 84.06 kg ha-1 K plus micronutrient followed by irrigation. On June 18,

2018 the plants were re-fertilized with 67.25 kg ha-1 N and 89.66 kg ha-1 K using the 6-

0-8 liquid fertilizer plus micronutrient package followed by irrigation.

Experimental Design

Completely randomized block design (RCBD) with 5 replications was used for the field experiment. Individuals of the same cross type were grouped together and randomized. Row to row distance was 1.22 m (4 ft) and plant to plant distance was maintained at 0.91 m (3ft).

Traits Evaluated

We measured yield related attributes of the PMN hybrids. Plant height, number of tillers, stem diameter, leaf length, leaf width, fresh biomass, and dry biomass were evaluated. For fresh biomass, all the tillers arising from individual plants were cut using a brush cutter and fresh weight measured with a hanging scale after seven months of planting. Samples for dry biomass were taken on August 20, 2018 by cutting 1-2 tillers

149

per plant for 3 replications of each line and weighed. They were then dried in a plant sample dryer at 48.8°C for 1 month after which the dry biomass weight of the samples was measured. The dry biomass for replicates was averaged and its derived function was used to extrapolate dry biomass from each line’s fresh biomass.

Data Analysis

Levene's test was done to compare the variances using 'car' package in R (Fox et al. 2012). Analysis of variance (ANOVA) and Tukey's HSD test were carried out using

'agricolae' package in R (De Mendiburu 2014). Linear model was fit using the trait value as dependent variable and cross type and replication as independent variable.

Results

Plant Height

Plant height of the four different crosses were significantly different to each other

(HSD, p<0.05). Variance due to block was significant which was accounted for in the model. The interaction between plant height and block was not significant. Cross D had significantly highest height followed by cross B, cross C, and cross A (Figure 5-2). The coefficient of variation of height was highest in cross type A (23.98%) and the lowest CV was on cross D (11.31%) (Table 5-2). Maximum height among all the groups was in cross D, with a height of 426.72 cm. Lowest range of plant height was in cross B

(228.60 cm – 396.24 cm) and the highest in cross D (86.36 cm – 426.72 cm) (Figure 5-

3).

Tiller Number

Tiller number varied significantly between the four types of crosses (p<0.05).

Block effect was not significant for tiller number. Tiller number was significantly higher in cross A followed by cross D (HSD, p<0.05) (Figure 5-4). Tiller number of cross B was at 150

par with cross C. Lowest number of tillers (1) was found in cross C and the highest number of tillers (43) was found in cross A. The CV of number of tillers was highest in cross A and lowest in cross D (Table 5-2). Similarly, the range of number of tillers was much greater in cross A (2 - 43) as compared to other crosses (3 – 22 in cross B, 1 – 27 in cross C, and 5 – 15 in cross D) (Figure 5-5).

Stem Diameter

Stem diameter between the groups were significantly different (p<0.05) and the block effect was not significant. Significantly highest (17.25 mm) stem diameter was found in cross B followed by cross C which was at par with cross D (Figure 5-6). Cross

A showed the highest CV (24.81%) for stem diameter while cross D showed the lowest

CV (14.16%) (Table 5-2). The range of stem diameter was wider in cross A (5.92 mm –

21.95 mm) compared to other crosses (10.04 – 25.71 mm for cross B, 9.63 mm – 22.65 mm for cross C, and 10.65 mm – 22.06 mm for cross D) (Figure 5-7).

Leaf Length

Leaf length was significantly different among the four crosses (p<0.05). Block effect was not significant. Leaf length of cross D was at par with that of cross B, while it was significantly higher than cross C or cross A (Figure 5-8). The minimum leaf length was found in cross A (82.25 cm) and the maximum leaf length was found in cross D

(106.84 cm). The CV for leaf length was highest in cross A (28.32%) and the lowest CV of leaf length was on cross D (10.15%) (Table 5-2). Range of leaf length was lower in cross D (72.39 cm – 132.08 cm) as compared to other crosses (38.10 cm – 130.81 cm for cross A, 26.67 cm – 129.54 cm for cross B, and 29.21 cm – 146.05 cm for cross D)

(Figure 5-9).

151

Leaf Width

Leaf width was significantly different among the four crosses (p<0.05). Block effect was not significant. Leaf width of cross D was at par with that of cross B, while it was significantly higher than cross C or cross A (Figure 5-10). On the other hand, leaf length of cross B was at par with cross D and cross C and significantly higher than cross A. The minimum leaf width was found in cross C (25 mm) and the maximum leaf width was found in cross B (69 mm). The CV for leaf width was highest in cross A

(18.80%) and the lowest CV of leaf width was on cross D (13.83%) (Table 5-2). Range of leaf width was lower in cross D (27 mm – 62 mm) as compared to other crosses (23 mm – 59 mm for cross A, 30 mm – 69 mm for cross B, and 25 mm – 64 mm for cross C)

(Figure 5-11).

Plant Biomass

Fresh biomass per plant (kg) was significantly different (p < 0.05) among the four crosses and the block effect was also significant. The mean response was different in block 4 and block 5 as compared to the other blocks and the block effect was accounted for in the model. The interaction between treatment and block was not significant.

Significantly highest biomass was found in cross D (8.54 kg/plant) followed by cross A

(6.88 kg/plant). Biomass for Cross C (4.97 kg/plant) and cross B (5.67 kg/plant) were at par with each other (Fig 5-12). The lowest biomass was found in cross B and the highest biomass was found in cross C (20.59 kg). The CV of biomass was lowest in cross D (33.63%) and highest in cross A (67.40%). Similarly, the spread of biomass values was wider in cross A (0.18 kg/plant – 19.16 kg/plant) as compared to other

152

crosses (1.56 kg/plant – 14.28 kg/plant in cross B, 0.12 kg/plant – 20.59 kg/plant in cross C and 3.06 kg/plant – 19.70 kg/plant in cross D) (Figure 5-13).

Dry Biomass

Dry biomass data from single plant was extrapolated to a larger land area in order to get dry biomass in tons ha-1. Projected dry biomass (tons ha-1) was significantly different (p < 0.05) among the four crosses. Significantly highest biomass was found in cross D (24.83 tons ha-1). Dry biomass for the remaining three cross types were at par with each other (18.39 tons ha-1, 18.34 tons ha-1, and 16.30 tons ha-1 for cross A, cross

B, and cross C, respectively) (Fig 5-14). Highest dry biomass was found in cross A

(54.87 tons ha-1) and the lowest was found in cross C (0.27 tons ha-1). The CV of dry biomass was lowest in cross D (33.63%) and highest in cross A (67.40%). Similarly, the range of biomass values among different progeny plants was much wider in cross A

(0.36 tons ha-1 - 54.51 tons ha-1) as compared to other crosses (3.84 tons ha-1 – 47.56 tons ha-1 in cross B, 0.27 tons ha-1 – 53.72 tons ha-1 in cross C, and 8.73 tons ha-1 –

48.59 tons ha-1 in cross D) (Figure 5-15).

Coefficient of Variation

The coefficient of variation (CV) for all measured traits is shown in Figure 5-16.

For all the traits evaluated, it was observed that the CV was least on cross D and highest on cross A. CV was the highest for dry biomass (tons/ha.) (30.95% - 73.42%)

(Table 5-2) and least for leaf width (13.83-18.80%) among all types of crosses. A relatively lower CV was observed for stem diameter (14.16-24.81%), plant height

(10.16-23.98%), and leaf length (10.15-28.32%). On the other hand, higher CV was

153

observed for number of tillers (31.93-48.47%). A picture of four different crosses during harvest is shown in Figure 5-17.

Correlation

All of the evaluated traits had significant, positive correlations with biomass weight. Correlation coefficient was the highest for number of tillers (0.60) and lowest for stem diameter (0.18). Correlation coefficients and p-values between biomass weight and each trait are presented in Table 5-3.

Discussion

Both pearl millet and napiergrass are protogynous and are predominantly cross- pollinated. This increases their heterozygosity which leads to a high level of heterosis in

PMN hybrids (Dowling, Burson, and Jessup 2014; J. B. Powell and Burton 1966).

Development of a cms pearl millet line optimized for biomass production will be the most efficient system for commercial scale production of large seeded PMN hybrids according to the method described by Powell and Burton (1966) (Dowling 2011). For commercial success of PMN hybrids, phenotypic stability and robustness of the genotype to consistently produce a specific phenotype is critical. To achieve this, we introgressed cms into high biomass pearl millet lines and assessed the phenotypic variation in biomass yield and related traits of PMN hybrids depending on different pearl millet sources. It should be noted that our attempt to produce a male napiergrass parent with higher level of homozygosity was not accomplished in time and therefore a heterozygous napiergrass parent was used in all these crosses. Inherent heterozygosity in this parent probably contributed to a high level of phenotypic variability in the progenies.

154

The biomass produced after seven months of growth was the highest in the cross involving the original pearl millet parent (cross D) followed by the cms parent, Tift 85

(cross A). The CV for biomass was also lowest in cross D (30.95%). However, CV for biomass of cross A was the highest (73.42%) as compared to cross B (39.66%) and cross C (56.55%). For the other traits evaluated, lowest CV was mostly found in cross D followed by cross B. As expected the high biomass pearl millet parent P787 used in cross D had a higher combining ability than the forage-type pearl millet parent Tift 85 resulting in significant differences in biomass yield. The backcross of Tift 85 with P787 was not able to restore the superior combining ability of P787. Interestingly, progenies from cross C using 787-S5 as parent showed a significant lower biomass yield than progenies from crosses involving the heterozygous P787. Testing the biomass accumulation of selfed progenies from P787 at every generation and selecting superior individuals may help to generate male sterile 787 with superior combining ability. The difference among CV of cross D and cross B was low. For plant height, the CV of cross

B was actually the lowest among the different crosses. CV for the various traits involving cross C (BC4 generation of Tift85 and 25-17) was lower than cross A but higher than cross B or cross D. We saw that the CV decreased in BC4 but hasn't reached the level of P787 or 787-S5. In theory, BC4 genotype will be 93.75% identical to the recurrent parent (Acquaah 2012). Therefore, in the line MS 787 BC4, we are in theory able to introgress approximately 94% of the genome of P787 parent. Continuing the back cross to more than BC7 generation will give 99% similarity to the P787 parent and may result in cms progenies that are more uniform.

155

Even though the average yield of PMN hybrids using cms parent Tift85 in cross A was high, the high variability in yield would make these lines less acceptable for commercial production because phenotypic uniformity contributes to a predictable yield.

Many factors such as variable genetic yield potential, uneven germination, variable planting depth, soil clods, insect damage, and moisture might be responsible for non- uniformity of plant stands etc. (Martin et al. 2005) which can be controlled by proper agronomic practices, experimental setup, and statistical analysis. Also, the cultivar resulting from back-crossing could differ from the initial cultivar beyond the transferred gene(s) because of linkage drag from the association of undesirable traits with the genes from the donor(Acquaah 2012).

Biomass yield of grasses is affected by different morphological traits. In sorghum, biomass is correlated with plant height, number of tillers, leaf length, leaf width, stem diameter, and flowering time (Hart et al. 2001; Murray et al. 2008; Xiao-ping et al. 2011).

Similarly, in Miscanthus, plant height, stem diameter, late flowering, and growth rate showed the highest positive correlation with yield (Zub, Arnoult, and Brancourt-Hulmel

2011). In Trichloris crinite, foliage height and basal diameter were strongly correlated with biomass yield (Cavagnaro et al. 2006). In napiergrass, plant height, number of tillers, and stem diameter were significantly correlated with plant biomass (Sinche

2013). Our results show that in PMN hybrids, traits like plant height, leaf length, leaf width, stem diameter, and number of tillers are significantly correlated with biomass yield. This indicates that tall PMN hybrids with numerous thick tillers and wide and long leaves produced more biomass than those with the opposite characteristics.

156

In this study, we have developed a cms version of elite pearl millet line P787.

The development of this non-dwarf biomass-type cms line of pearl millet provides new resources for pearl millet and napiergrass breeding that can be exploited for commercial production of uniform PMN hybrids.

157

Figure 5-1. Seeds of pearl millet dwarf cms line (A), cms high biomass pearl millet line (B), PMN hybrid (C), and napiergrass (D), respectively from left to right.

158

400

300

Plant height (cm) Plantheight 200

100 d b c a

A B C D Cross Type

Figure 5-2. Boxplot of plant height (cm) for four different types of crosses studied. Small colored dots represent individual plant height of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

159

A B

C D

0 100 200 300 400 100 200 300 400 Plant height (cm)

Figure 5-3. Histogram of plant height (cm) for four different crosses. X-axis represent plant height in cm and y-axis represents the count of plants. A = Tift 85 × 25- 17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

160

20 Number of tillers of Number

a c c b 0 A B C D Cross Type

Figure 5-4. Boxplot of number of tillers for four different types of crosses studied. Small colored dots represent number of tillers for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787- S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

161

A B

C D

0 0 10 20 30 40 0 10 20 30 40 Number of tillers

Figure 5-5. Histogram of number of tillers for four different crosses. X-axis represent the number of tillers and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

162

15 Stem diameter (mm) diameter Stem

c a b b 5 A B C D Cross Type

Figure 5-6. Boxplot of stem diameter (mm) for four different types of crosses studied. Small colored dots represent the stem diameter (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787- S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

163

A B

C D

0 5 10 15 20 25 5 10 15 20 25 Stem diameter (mm)

Figure 5-7. Histogram of stem diameter (mm) for four different crosses. X-axis represent the stem diameter (mm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

164

150

100 Leaf length (cm) length Leaf

c a b a

A B C D Cross Type

Figure 5-8. Boxplot of leaf length (cm) for four different types of crosses studied. Small colored dots represent the leaf length (cm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787- S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

165

A B 40

C D 40

0 50 100 150 50 100 150 Leaf length (cm)

Figure 5-9. Histogram of leaf length (cm) for four different crosses. X-axis represent the leaf length (cm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

166

40 Leaf width(mm) Leaf

c ab b a

A B C D Cross Type

Figure 5-10. Boxplot of leaf width (mm) for four different types of crosses studied. Small colored dots represent the leaf width (mm) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25- 17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

167

A B

C D

0 30 40 50 60 70 30 40 50 60 70 Leaf width (mm)

Figure 5-11. Histogram of leaf width (mm) for four different crosses. X-axis represent the leaf width (mm) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

168

Fresh Biomass(kg/plant) Fresh 5

0 b c c a

A B C D Cross Type

Figure 5-12. Boxplot of fresh biomass per plant (kg) for four different types of crosses studied. Small colored dots represent the fresh biomass per plant (kg) for individual plants of the cross. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

169

A B

C D

0 0 5 10 15 20 0 5 10 15 20 Fresh Biomass (kg/plant)

Figure 5-13. Histogram of fresh biomass per plant (kg) for four different crosses. X-axis represent the fresh biomass per plant (kg) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

170

20 Dry Biomass(tons/ha.) Dry

0 b b b a

A B C D Cross Type

Figure 5-14. Boxplot of projected dry biomass (tons per ha) for four different types of crosses evaluated at seven months of growth. Different letters in blue color indicate significant differences among crosses based on Tukey’s HSD test at p ≤ 0.05. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

171

A B 20

C D 20

0 0 20 40 0 20 40 Dry Biomass (tons/ha)

Figure 5-15. Histogram of dry biomass (tons per ha.) for four different crosses evaluated at seven months of growth. X-axis represent the dry biomass (tons per ha.) and y-axis represents the count of plants. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

172

BioMass height LeafLength

LeafWidth StemDiam TillerNum

60 Coefficient of variation (%) variation of Coefficient

0 A B C D A B C D A B C D Cross type Figure 5-16. Coefficient of variation (%) for the different traits evaluated. A = Tift 85 × 25-17, B = 787-S5 × 25-17, C = MS 787 BC4 × 25-17, D = P787 × 25-17.

173

Figure 5-17. Picture of four different crosses during harvest in Citra, FL. Cross A = Tift 85 × 25-17, Cross B = 787-S5 × 25-17, Cross C = MS 787 BC4 × 25-17, Cross D = P787 × 25-17.

174

Table 5-1. Details of the cross types used in the experiment. Cross type Female parent ♀ Male parent ♂ Cross details Number of (Pearl millet) (Napiergrass) plants A Tift85 × 25-17 cms parent 40 B 787-S5 × 25-17 Selfed 5th gen. 50 C MS 787 BC4 × 25-17 BC4 generation 27 D P787 × 25-17 Original parent 22

175

Table 5-2. Descriptive statistics for different types of crosses. P-value represents Pr>F for Levene’s test for homogeneity of variance (center = median), CV = coefficient of variation. Trait Cross type P-value Mean Standard Deviation CV (%) Min Max

Plant height 0.0000

Plant height A 0.0778 223.84 53.67 23.98 86.36 340.36

Plant height B 0.9584 313.41 31.84 10.16 228.60 396.24

Plant height C 0.8109 285.09 50.58 17.74 111.76 375.92

Plant height D 0.6924 328.32 37.13 11.31 86.36 426.72

Number of tillers 0.0000

Number of tillers A 0.0758 17.88 8.67 48.47 2.00 43.00

Number of tillers B 0.6613 9.46 3.69 39.03 3.00 22.00

Number of tillers C 0.1015 8.98 4.29 47.73 1.00 27.00

Number of tillers D 0.6265 12.82 4.09 31.93 5.00 15.00

Stem diameter 0.0000

Stem diameter A 0.3174 13.31 3.30 24.81 5.92 21.95

Stem diameter B 0.9829 17.26 2.73 15.80 10.04 25.71

Stem diameter C 0.4913 16.42 2.61 15.90 9.63 22.65

Stem diameter D 0.8523 16.41 2.32 14.16 10.65 22.06

Leaf length 0.0000

Leaf length A 0.2382 82.26 23.29 28.32 38.10 130.81

Leaf length B 0.5612 106.01 11.75 11.08 26.67 129.54

Leaf length C 0.6267 101.17 15.05 14.88 29.21 146.05

Leaf length D 0.1045 106.84 10.84 10.15 72.39 132.08

Leaf width 0.0529

Leaf width A 0.4596 40.75 7.66 18.80 23.00 59.00

Leaf width B 0.8497 47.61 6.88 14.46 30.00 69.00

Leaf width C 0.7394 46.04 7.92 17.20 25.00 64.00

176

Table 5-2. Continued Trait Cross type P- Mean Standard CV (%) Min Max value Deviation

Leaf width D 0.1875 48.89 6.76 13.83 27.00 62.00

Fresh biomass 0.0000

Fresh biomass A 0.0214* 6.88 4.64 67.40 0.18 19.16

Fresh biomass B 0.6081 6.23 2.29 36.75 1.56 14.28

Fresh biomass C 0.0794 5.47 3.10 56.63 0.12 20.59

Fresh biomass D 0.7515 8.54 2.87 33.63 3.06 19.70

Dry biomass 0.0000

Dry biomass A 0.0207* 18.39 13.50 73.42 0.36 54.87

Dry biomass B 0.6462 18.34 7.27 39.66 3.84 47.56

Dry biomass C 0.1246 16.30 9.22 56.55 0.27 53.72

Dry biomass D 0.9255 24.83 7.68 30.95 8.73 48.59

177

Table 5-3. Correlation coefficients and p-values for biomass weight and biomass- related traits for PMN hybrids evaluated in Citra, FL. Plant height Leaf length Leaf width Stem diameter Number of tillers R 0.43 0.36 0.25 0.18 0.6 p-value <0.001 <0.001 <0.001 <0.001 <0.001

178

CHAPTER 6 CONCLUDING REMARKS

Summary

Napiergrass is an important forage and biofuel crop. However, its commercial utilization has been lagging behind other crops due to limited genetic and genomic resources. Before this study, there were no SSR markers derived from napiergrass sequences publicly available, no genetic linkage map was available, and no QTL studies for any traits were published. The lack of these genetic resources limits the advances that can be achieved via breeding. In this project, we have developed genetic and genomic resources that will be important in napiergrass breeding. In Chapter 2, we constructed the first high-density genetic linkage map of napiergrass using NGS-derived

SNP markers. We also identified 5,339 SSRs using napiergrass genomic sequences and successfully designed primers for 1,926 SSRs. These results will be useful for future molecular breeding programs such as identification of QTLs for important traits as well as MAS for the genetic improvement of napiergrass and comparative genomics.

Early flowering cultivars of napiergrass produce abundant wind dispersed seeds, which contribute to high potential of invasiveness. Controlling flowering or modifying flowering time of napiergrass can help in reducing its invasiveness and boost its potential as biofuel feedstock. Therefore, a better understanding of the genetic basis of flowering time in napiergrass is necessary. To facilitate this, in Chapter 3, we demonstrated that flowering time and number of flowers are highly heritable traits.

Therefore, we conducted the first QTL analysis in napiergrass and identified three stable QTLs controlling number of flowers. We also identified three potential QTLs that

179

control flowering time in napiergrass. We were able to identify potential candidate genes such as AGAMOUS, DEFICIENS, DELLA, WRKY, and SERK1 that were harbored by the QTL regions. The QTLs detected in this study will be valuable tools for napiergrass breeding and marker assisted selection. Similarly, the candidate genes identified could be potential targets for genome editing to modify flowering time in napiergrass.

To improve our understanding of the genetic basis of flowering time in napiergrass and to evaluate the genetic diversity of napiergrass, we used exome- capture sequencing to study the germplasm collection of napiergrass, described in

Chapter 4. We identified 78,129 high quality SNPs in the germplasm collection that serve as novel genomic resources for napiergrass breeding programs. We also identified potential candidate genes like Calcium-binding protein CML, enoyl-CoA hydratase, phytochrome B, and AP2-like factors in the germplasm collection. This study showed the feasibility to apply targeted exome sequencing with probes designed from

CDS of pearl millet to target the genomic regions in napiergrass.

We also used traditional breeding approaches to increase biosafety of napiergrass. For this, in Chapter 5, we introgressed cms available in forage type dwarf pearl millet lines to high biomass yielding pearl millet lines. We then hybridized these elite cms lines of pearl millet with napiergrass to generate PMN hybrids that are male and female sterile and will not contribute to invasiveness with wind dispersed seeds. We studied the uniformity of different types of parental combinations. Substantial variation within each cross was found for biomass accumulation and yield related traits.

Uniformity of the progenies could be enhanced by using homozygous parents to make the crosses. If efficient seed production can be accomplished under field conditions and

180

if uniformity can be improved without compromising biomass yield, PMN hybrids may outperform alternative forage and biofuel crops in the near future.

Future work

The construction of the first genetic map of napiergrass has opened up new avenues in napiergrass breeding. With this map, we were able to identify QTLs related to flowering time and number of flowers in napiergrass. There are other traits like stem diameter, tiller number, and yield, whose genetic basis needs to be further explored. It is our hope that the genetic map developed in Chapter 2 can be utilized to explore other traits of agronomic importance to identify QTLs and candidate genes for these traits.

Further, the sequence variations identified in the germplasm collection can be utilized in the future to study the genetic architecture of other traits and to perform GWAS for these traits. Moreover, introgression of cms into high biomass lines like P787 will continue to obtain a homozygous cms line with superior combining ability with napiergrass. Similarly, generation of homozygous or near-homozygous napiergrass lines will have to be continued by repeated self-fertilization and selection of the best performing lines. PMN hybrids from these homozygous or near homozygous parents should be evaluated not only for increased uniformity but also for biomass yield which may mainly depend on the different level of heterosis in the A and B genomes of the hybrids and the combining ability of the parents. For commercial production of PMN hybrid seeds under field condition different accessions need to be evaluated to synchronize flowering time and identify parents that contribute to the highest seed yield.

Alternatively, highest yielding napiergrass lines can be evaluated in countries where

181

napiergrass is a native species, the preferred forage crop and invasiveness is not a concern.

182

LIST OF REFERENCES

Acquaah, George. 2012. “Breeding Self Pollinated Species.” In Principles of Plant Genetics and Breeding, Second Edition, 303–36. Wiley Online Library.

Adhikari, Laxman, Orville M Lindstrom, Jonathan Markham, and Ali M Missaoui. 2018. “Dissecting Key Adaptation Traits in the Polyploid Perennial Medicago Sativa Using GBS-SNP Mapping.” Frontiers in Microbiology 9: 1–19. doi:10.3389/fpls.2018.00934.

Agarwal, Milee, Neeta Shrivastava, and Harish Padh. 2008. “Advances in Molecular Marker Techniques and Their Applications in Plant Sciences.” Plant Cell Reports 27 (4): 617–31. doi:10.1007/s00299-008-0507-z.

Aken’ova, M. E., and H. R. Chheda. 1973. “Interspecific Hybrids Pennisetum Typhoides S & H × P. Purpureum Schum. for Forage in the Hot Humid Lowland Tropics of West Africa.” The Nigerian Agricultural Journal 10 (10): 82–90.

Andersen, Jeppe R, and Thomas Lübberstedt. 2003. “Functional Markers in Plants.” Trends in Plant Science 8 (11): 554–60. doi:10.1016/j.tplants.2003.09.010.

Anderson, William F., Michael Casler, and Brian Baldwin. 2008. “Improvement of Perennial Forage Species as Feedstock for Bioenergy.” In Genetic Improvement of Bioenergy Crops, edited by Wilfred Vermerris, 309–45. Springer Science+Business Media, LLC.

Atienza, S. G., Z. Satovic, K. K. Petersen, O. Dolstra, and A. Martín. 2003. “Identification of QTLs Influencing Combustion Quality in Miscanthus Sinensis Anderss. II. Chlorine and Potassium Content.” Theoretical and Applied Genetics, no. 107: 857–63. doi:10.1007/s00122-003-1218-z.

Augustin, Eiiane, and Fernando Adami Tcacenco. 1993. “Isoenzymatic Characterization of Elephantgrass (Pennisetum Purpureum Schum.) Germplasm.” Rev. Brasil. Genet. 16 (3): 685–96.

Azevedo, Ana Luisa Sousa, Pricila Palla Costa, Juarez Campolina Machado, Marco Antonio Machado, Antônio Vander Pereira, and Francisco José da Silva Lédo. 2012. “Cross Species Amplification of Pennisetum Glaucum Microsatellite Markers in Pennisetum Purpureum and Genetic Diversity of Napier Grass Accessions.” Crop Science 52: 1776–85. doi:10.2135/cropsci2011.09.0480.

Bastow, Ruth, and Caroline Dean. 2002. “The Molecular Basis of Photoperiodism.” Biological Rhythm Research 37 (4): 353–80. doi:10.1016/S1534-5807(02)00296- 4.

183

Bezant, Jeremy, David A. Laurie, Nicky Pratchett, Jan A N Chojecki, and Mike J. Kearsey. 1996. “Marker Regression Mapping of QTL Controlling Flowering Time and Plant Height in a Spring Barley (Hordeum Vulgare L .) Cross.” Heredity 77: 64–73.

Bhandari, Anitha P, D H Sukanya, and C R Ramesh. 2006. “Application of Isozyme Data in Fingerprinting Napier Grass (Pennisetum Purpureum Schum.) for Germplasm Management.” Genetic Resources and Crop Evolution 53. Springer: 253–64.

Bidinger, F. R., T. Nepolean, C. T. Hash, R. S. Yadav, and C. J. Howarth. 2007. “Quantitative Trait Loci for Grain Yield in Pearl Millet under Variable Postflowering Moisture Conditions.” Crop Science 47 (3): 969–80. doi:10.2135/cropsci2006.07.0465.

Bielenberg, Douglas Gary, Bradley Rauh, Shenghua Fan, Ksenija Gasic, Albert Glenn Abbott, Gregory Lynn Reighard, William R. Okie, and Christina Elizabeth Wells. 2015. “Genotyping by Sequencing for SNP-Based Linkage Map Construction and QTL Analysis of Chilling Requirement and Bloom Date in Peach [Prunus Persica (L.) Batsch].” PLoS ONE, 1–14. doi:10.1371/journal.pone.0139406.

Bodénès, Catherine, Emilie Chancerel, François Ehrenmann, Antoine Kremer, and Christophe Plomion. 2016. “High-Density Linkage Mapping and Distribution of Segregation Distortion Regions in the Oak Genome.” DNA Research 23 (2): 115–24. doi:10.1093/dnares/dsw001.

Bohuon, E J R, L D Ramsay, J A Craft, A E Arthur, D F Marshall, D J Lydiate, and Mike J. Kearsey. 1998. “The Association of Flowering Time Quantitative Trait Loci with Duplicated Regions and Candidate Loci in Brassica Oleracea.” Genetics 150: 393–401.

Bolger, Anthony M., Marc Lohse, and Bjoern Usadel. 2014. “Trimmomatic: A Flexible Trimmer for Illumina Sequence Data.” Bioinformatics 30 (15): 2114–20. doi:10.1093/bioinformatics/btu170.

Boonman, Joseph G. 1988. East Africa’s Grasses and Fodders: Their Ecology and Husbandry. Edited by Helmut Lieth and Harold A. Mooney. Springer- Science+Business Media, B.V.

Bouché, Frédéric, Guillaume Lobet, Pierre Tocquin, and Claire Périlleux. 2016. “FLOR- ID: An Interactive Database of Flowering-Time Gene Networks in Arabidopsis Thaliana.” Nucleic Acids Research 44: D1167–71. doi:10.1093/nar/gkv1054.

Bourke, Peter M., Roeland E. Voorrips, Richard G. F. Visser, and Chris Maliepaard. 2018. “Tools for Genetic Studies in Experimental Populations of Polyploids.” Frontiers in Plant Science 9: 1–17. doi:10.3389/fpls.2018.00513.

184

Brunken, Jere N. 1977. “A Systematic Study of Pennisetum Sect . Pennisetum (Gramineae).” American Journal of Botany 64 (2): 161–76.

Buckler, Edward S, James B Holland, Peter J Bradbury, Charlotte B Acharya, Patrick J Brown, Chris Browne, Elhan Ersoz, Sherry Flint-Garcia, Arturo Garcia, and Jeffrey C. Glaubitz. 2009. “The Genetic Architecture of Maize Flowering Time.” Science 325. American Association for the Advancement of Science: 714–18.

Burton, Glenn W. 1944. “Hybrids between Napier Grass and Cattail Millet.” The Journal of Heredity, 227–32.

———. 1958. “Cytoplasmic Male-Sterility in Pearl Millet (Pennisetum glaucum)(L.) R. Br.” Agronomy Journal 50 (4). American Society of Agronomy: 230.

———. 1990. Grasses : New and Improved. Edited by J. Janick and J.E. Simon. Advances in New Crops. Portland, Oregon: Timber Press.

Burton, Glenn W., and D S Athwal. 1967. “Two Additional Sources of Cytoplasmic Male- Sterility in Pearl Millet and Their Relationship to Tift 23A.” Crop Science 7 (3). Crop Science Society of America: 209–11.

Cai, C, F Y Cheng, J Wu, Y Zhong, and G Liu. 2015. “The First High-Density Genetic Map Construction in Tree Peony (Paeonia Sect. Moutan) Using Genotyping by Specific-Locus Amplified Fragment Sequencing.” PLoS One, 1–16. doi:10.1371/journal.pone.0128584.

Catchen, Julian, A. Amores, P. Hohenlohe, W. Cresko, J. H. Postlethwait, and D.-J. De Koning. 2011. “Stacks: Building and Genotyping Loci de Novo from Short-Read Sequences.” G3.Genes|Genomes|Genetics 1: 171–82. doi:10.1534/g3.111.000240.

Catchen, Julian, Paul A. Hohenlohe, Susan Bassham, Angel Amores, and William A. Cresko. 2013. “Stacks: An Analysis Tool Set for Population Genomics.” Molecular Ecology 22: 3124–40. doi:10.1111/mec.12354.

Cavagnaro, Pablo F, Juan B Cavagnaro, José L Lemes, Ricardo W Masuelli, and Carlos B Passera. 2006. “Genetic Diversity among Varieties of the Native Forage Grass Trichloris Crinita Based on AFLP Markers, Morphological Characters, and Quantitative Agronomic Traits.” Genome 49: 906–18. doi:10.1139/g06-060.

Chemisquy, M. Amelia, Liliana M. Giussani, María A. Scataglini, Elizabeth A. Kellogg, and Osvaldo Morrone. 2010. “Phylogenetic Studies Favour the Unification of Pennisetum, Cenchrus and Odontelytrum (Poaceae): A Combined Nuclear, Plastid and Morphological Analysis, and Nomenclatural Combinations in Cenchrus.” Annals of Botany 106: 107–30. doi:10.1093/aob/mcq090.

185

Churchill, Gary A., and R. W. Doerge. 1994. “Empirical Threshold Values for Quantitative Trait Mapping.” Genetics 138: 963–71. doi:10.1534/genetics.107.080101.

Cingolani, Pablo, Adrian Platts, Le Lily Luan Wang, Melissa Coon, Tung Nguyen, Le Lily Luan Wang, Susan J. Land, Xiangyi Lu, and Douglas M. Ruden. 2012. “A Program for Annotating and Predicting the Effects of Single Nucleotide Polymorphisms, SnpEff.” Fly 6 (2): 80–92.

Clevenger, Josh, Carolina Chavarro, Stephanie A. Pearl, Peggy Ozias-Akins, and Scott A. Jackson. 2015. “Single Nucleotide Polymorphism Identification in Polyploids: A Review, Example, and Recommendations.” Molecular Plant 8 (6). Elsevier Ltd: 831–46. doi:10.1016/j.molp.2015.02.002.

Crawford, Joseph, Patrick J. Brown, Thomas Voigt, and D. K. Lee. 2016. “Linkage Mapping in Prairie Cordgrass (Spartina Pectinata Link) Using Genotyping-by- Sequencing.” Molecular Breeding 36 (5). Springer Netherlands: 1–12. doi:10.1007/s11032-016-0484-9.

Cuomo, Greg J, David C Blouin, and James E Beatty. 1996. “Forage Potential of Dwarf Napiergrass and a Pearl Millet X Napiergrass Hybrid.” Agronomy Journal 88: 434–38.

D’Antonio, Carla M, and Peter M Vitousek. 1992. “Biological Invasions by Exotic Grasses, the Grass/fire Cycle, and Global Change.” Annual Review of Ecology and Systematics 23. JSTOR: 63–87. de Lima, R S N, R F Daher, L S a Gonçalves, D a Rossi, a T do Amaral Júnior, M G Pereira, and F J S Lédo. 2011. “RAPD and ISSR Markers in the Evaluation of Genetic Divergence among Accessions of Elephant Grass.” Genetics and Molecular Research : GMR 10 (3): 1304–13. doi:10.4238/vol10-3gmr1107.

De Mendiburu, Felipe. 2014. “Agricolae: Statistical Procedures for Agricultural Research.” R Package Version.

Dell’Acqua, Matteo, Daniel M. Gatti, Giorgio Pea, Federica Cattonaro, Frederik Coppens, Gabriele Magris, Aye L. Hlaing, et al. 2015. “Genetic Properties of the MAGIC Maize Population: A New Platform for High Definition QTL Mapping in Zea Mays.” Genome Biology 16 (167). Genome Biology: 1–23. doi:10.1186/s13059-015-0716-z.

Dellaporta, Stephen L, Jonathan Wood, and James B Hicks. 1983. “A Plant DNA Minipreparation: Version II.” Plant Molecular Biology Reporter 1 (4). Springer: 19–21.

186

Deschamps, Stéphane, Victor Llaca, and Gregory D. May. 2012. “Genotyping-by- Sequencing in Plants.” Biology 1. Molecular Diversity Preservation International: 460–83. doi:10.3390/biology1030460.

Devos, Katrien M, and Mike D Gale. 2000. “Genome Relationships: The Grass Model in Current Research.” The Plant Cell 12. Am Soc Plant Biol: 637–46.

Devos, Katrien M, T S Pittaway, A Reynolds, and M D Gale. 2000. “Comparative Mapping Reveals a Complex Relationship between the Pearl Millet Genome and Those of Foxtail Millet and Rice.” Theoretical and Applied Genetics 100. Springer: 190–98. doi:10.1007/s001220050026.

Diz, D. A., and S. C. Schank. 1993. “Characterization of Seed Producing Pearl Millet X Elephantgrass Hexaploid Hybrids.” Euphytica 67: 143–49. doi:10.1007/BF00022737.

Dowling, Charlie D. 2011. “Marker-Assisted Verification of Hybrids in Pearl Millet- Napiergrass (Pennisetum Glaucum [L.] R. BR. X Pennisetum Purpureum Schumach.).” Texas A&M University.

Dowling, Charlie D., Byron L Burson, Jamie L Foster, Lee Tarpley, and Russell W Jessup. 2013. “Confirmation of Pearl Millet-Napiergrass Hybrids Using EST- Derived Simple Sequence Repeat (SSR) Markers.” American Journal of Plant Sciences 4. Scientific Research Publishing: 1004–12.

Dowling, Charlie D., Byron L Burson, and Russell W Jessup. 2014. “Marker-Assisted Verification of Kinggrass (Pennisetum Purpureum Schumach . X Pennisetum Glaucum [ L .] R . Br .).” Plant Omics Journal 7 (2): 72–79.

Durand, Eléonore, Sophie Bouchet, Pascal Bertin, Adrienne Ressayre, Philippe Jamin, Alain Charcosset, Christine Dillmann, and Maud I Tenaillon. 2012. “Flowering Time in Maize: Linkage and Epistasis at a Major Effect Locus.” Genetics 190. Genetics Soc America: 1547–62.

Elshire, Robert J., Jeffrey C. Glaubitz, Qi Sun, Jesse A. Poland, Ken Kawamoto, Edward S. Buckler, and Sharon E. Mitchell. 2011. “A Robust, Simple Genotyping- by-Sequencing (GBS) Approach for High Diversity Species.” PLoS ONE 6 (5). doi:10.1371/journal.pone.0019379.

Erickson, John E., Arkorn Soikaew, Lynn E. Sollenberger, and Jerry M. Bennett. 2012. “Water Use and Water-Use Efficiency of Three Perennial Bioenergy Grass Crops in Florida.” Agriculture 2: 325–38. doi:10.3390/agriculture2040325.

Falconer, Douglas S, Trudy F C Mackay, and Richard Frankham. 1996. “Introduction to Quantitative Genetics.” In Trends in Genetics, 4th ed. [Amsterdam, The Netherlands: Elsevier Science Publishers (Biomedical Division)], c1985-.

187

Faris, J. D., B. Laddomada, and B. S. Gill. 1998. “Molecular Mapping of Segregation Distortion Loci in Aegilops Tauschii.” Genetics 149 (1): 319–27.

Farrell, G., S. A. Simons, and R. J. Hillocks. 2002. “Pests, Diseases and Weeds of Napier Grass, Pennisetum Purpureum: A Review.” International Journal of Pest Management 48 (1). Taylor & Francis: 39–48. doi:10.1080/09670870110065578.

Ferrão, Luís Felipe V., Juliana Benevenuto, Ivone de Bem Oliveira, Catherine Cellon, James Olmstead, Matias Kirst, Marcio F. R. Resende, and Patricio Munoz. 2018. “Insights into the Genetic Basis of Blueberry Fruit-Related Traits Using Diploid and Polyploid Models in a GWAS Context.” Frontiers in Ecology and Evolution 6. doi:10.3389/fevo.2018.00107.

FLEPPC. 2011. “List of Invasive Plant Species.” Florida Exotic Pest Plant Council. http://www.fleppc.org/list/2011plantlist.pdf.

Fox, John, Sanford Weisberg, Daniel Adler, Douglas Bates, Gabriel Baud-Bovy, Steve Ellison, David Firth, Michael Friendly, Gregor Gorjanc, and Spencer Graves. 2012. “Package ‘car.’” Vienna: R Foundation for Statistical Computing.

Fraser, Hunter B., and Eric E. Schadt. 2010. “The Quantitative Genetics of Phenotypic Robustness.” PLoS ONE 5 (1). doi:10.1371/journal.pone.0008635.

Ganal, Martin W, Thomas Altmann, and Marion S Röder. 2009. “SNP Identification in Crop Plants.” Current Opinion in Plant Biology 12: 211–17. doi:10.1016/j.pbi.2008.12.009.

Garner, Wightman Wells. 1933. “Comparative Responses of Long-Day and Short-Day Plants to Relative Length of Day and Night.” Plant Physiology 8 (3). American Society of Plant Biologists: 347–56.

Garrison, Erik, and Gabor Marth. 2012. “Haplotype-Based Variant Detection from Short- Read Sequencing.” arXiv:1207.3907.

Gelli, Malleswari, Sharon E. Mitchell, Kan Liu, Thomas E. Clemente, Donald P. Weeks, Chi Zhang, David R. Holding, and Ismail M. Dweikat. 2016. “Mapping QTLs and Association of Differentially Expressed Gene Transcripts for Multiple Agronomic Traits under Different Nitrogen Levels in Sorghum.” BMC Plant Biology 16 (16). BMC Plant Biology: 1–18. doi:10.1186/s12870-015-0696-x.

Gifford, Justin M., Won Byong Chae, Kankshita Swaminathan, Stephen P. Moose, and John A. Juvik. 2015. “Mapping the Genome of Miscanthus Sinensis for QTL Associated with Biomass Productivity.” Global Change Biology Bioenergy 7: 797–810. doi:10.1111/gcbb.12201.

188

Glaubitz, Jeffrey C., Terry M. Casstevens, Fei Lu, James Harriman, Robert J. Elshire, Qi Sun, and Edward S. Buckler. 2014. “TASSEL-GBS: A High Capacity Genotyping by Sequencing Analysis Pipeline.” PLoS ONE 9 (2). doi:10.1371/journal.pone.0090346.

Gonzalez, Boldomero, and Wayne W. Hanna. 1984. “Morphological and Fertility Responses in Isogenic Triploid and Hexaploid Pearl Millet X Napiergrass Hybrids.” The Journal of Heredity 75: 317–18.

Grabowski, Paul P., Joseph Evans, Chris Daum, Shweta Deshpande, Kerrie W. Barry, Megan Kennedy, Guillaume Ramstein, et al. 2016. “Genome-Wide Associations with Flowering Time in Switchgrass Using Exome-Capture Sequencing Data.” New Phytologist. doi:10.1111/nph.14101.

Gupta, S C, and O Mhere. 1997. “Identification of Superior Pearl Millet by Napier Hybrids and Napiers in Zimbabwe.” African Crop Science Journal 5 (3). African Crop Science Society: 229–37.

Haggard, J. Erron, Emily B. Johnson, and Dina A. St. Clair. 2015. “Multiple QTL for Horticultural Traits and Quantitative Resistance to Phytophthora Infestans Linked on Solanum Habrochaites Chromosome 11.” G3.Genes|Genomes|Genetics 5: 219–33. doi:10.1534/g3.114.014654.

Hanna, Wayne W. 1981. “Method of Reproduction in Napiergrass and in the 3X and 6X Alloploid Hybrids with Pearl Millet.” Crop Science. Crop Science Society of America, 123–26.

———. 1987. “Utilization of Wild Relatives of Pearl Millet.” In Proceedings of the International Pearl Millet Workshop, 33–42. Patancheru, India: ICRISAT.

Hanna, Wayne W., and Warren G Monson. 1980. “Yield, Quality, and Breeding Behavior of Pearl Millet ✕ Napiergrass Interspecific Hybrids.” Agronomy Journal 72 (2). American Society of Agronomy: 358–60.

Hanna, Wayne W., H D Wells, and Glenn W. Burton. 1987. “Registration of Pearl Millet Inbred Parental Lines, Tift 85D2A1 and Tift 85D2B1.” Crop Science 27: 1324–25.

Harlan, J R, and J M J de Wet. 1971. “Toward a Rational Classification of Cultivated Plants.” Taxon 20 (4). International Association for Plant Taxonomy (IAPT): 509– 17. doi:10.2307/1218252.

Harris-Shultz, Karen R., William F. Anderson, and Ravindra Malik. 2010. “Genetic Relationships among Napiergrass (Pennisetum Purpureum Schum.) Nursery Accessions Using AFLP Markers.” Plant Genetic Resources: Characterization and Utilization 8 (1). Cambridge Univ Press: 63–70.

189

Hart, G. E., K. F. Schertz, Y. Peng, and N. H. Syed. 2001. “Genetic Mapping of Sorghum Bicolor (L.) Moench QTLs That Control Variation in Tillering and Other Morphological Characters.” Theoretical and Applied Genetics 103: 1232–42. doi:10.1007/s001220100582.

Havey, Michael J. 2004. “The Use of Cytoplasmic Male Sterility for Hybrid Seed Production.” In Molecular Biology and Biotechnology of Plant Organelles, edited by H. Daniell and C. D. Chase, 623–34. doi:10.1007/978-1-4020-3166-3_23.

Huang, Xiaoen, Fangfang Wang, Ratnesh Singh, James A Reinert, M C Engelke, Anthony D Genovesi, Ambika Chandra, and Qingyi Yu. 2016. “Construction of High-Resolution Genetic Maps of Zoysia Matrella ( L .) Merrill and Applications to Comparative Genomic Analysis and QTL Mapping of Resistance to Fall Armyworm.” BMC Genomics 17 (562): 1–16. doi:10.1186/s12864-016-2969-7.

Huang, Xiaoqiu, and Anup Madan. 1999. “CAP3: A DNA Sequence Assembly Program.” Genome Research 9: 868–77. doi:10.1101/gr.9.9.868.

Hussain, Waseem, P. Stephen Baenziger, Vikas Belamkar, Mary J. Guttieri, Jorge P. Venegas, Amanda Easterly, Ahmed Sallam, and Jesse A. Poland. 2017. “Genotyping-by-Sequencing Derived High-Density Linkage Map and Its Application to QTL Mapping of Flag Leaf Traits in Bread Wheat.” Scientific Reports 7. Springer US: 1–15. doi:10.1038/s41598-017-16006-z.

ICRISAT, and FAO. 1996. The World Sorghum and Millet Economies: Facts, Trends and Outlook. Patancheru, India: International Crops Research Institute for the Semi-Arid Tropics, and Rome, Italy: Food and Agricultural Organization of the United Nations.

Jauhar, Prem P. 1981. Cytogenetics and Breeding of Pearl Millet and Related Species. Alan R. Liss, Inc. Progress and Topics in Cytogenetics, Volume 1.

Jauhar, Prem P, and Wayne W. Hanna. 1998. Cytogenetics and Genetics of Pearl Millet. Edited by Donald L. Sparks. Advances in Agronomy. Vol. 64. New York: Academic Press, 1949-. doi:10.1017/CBO9781107415324.004.

Jessup, Russell W. 2013. “‘Seeded-yet-Sterile’ Perennial Biofuel Feedstocks.” Advances in Crop Science and Technology 1 (2): 1–2.

Jiang, Guo-liang. 2013. “Molecular Markers and Marker-Assisted Breeding in Plants.” In Plant Breeding from Laboratories to Fields, 45–83. InTech.

Jombart, Thibaut. 2008. “Adegenet: A R Package for the Multivariate Analysis of Genetic Markers.” Bioinformatics 24 (11): 1403–5. doi:10.1093/bioinformatics/btn129.

190

Juma, Awalla Borniface. 2014. “Genetic Diversity and Population Structure of Napier Grass (Pennisetum Purpureum Schumach) in Western Kenya.” University of Eldoret, Kenya.

Jung, Christian, Gina Capistrano-Gossmann, Janina Braatz, Niharika Sashidhar, and Siegbert Melzer. 2018. “Recent Developments in Genome Editing and Applications in Plant Breeding.” Plant Breeding 137 (1): 1–9. doi:10.1111/pbr.12526.

Jung, Christian, and Andreas E Müller. 2009. “Flowering Time Control and Applications in Plant Breeding.” Trends in Plant Science 14 (10): 563–73. doi:10.1016/j.tplants.2009.07.005.

Kandel, Raju, Hari P. Singh, Bharat P. Singh, Karen R. Harris-Shultz, and William F. Anderson. 2016. “Assessment of Genetic Diversity in Napier Grass (Pennisetum Purpureum Schum.) Using Microsatellite, Single-Nucleotide Polymorphism and Insertion-Deletion Markers from Pearl Millet (Pennisetum Glaucum [L.] R. Br.).” Plant Molecular Biology Reporter 34: 265–72. doi:10.1007/s11105-015-0918-2.

Kannan, Baskaran, Senthilvel Senapathy, Arcot Gajaraj Bhasker Raj, Subhash Chandra, Arunachalam Muthiah, Arun Prabhu Dhanapal, and C Thomas Hash. 2014. “Association Analysis of SSR Markers with Phenology, Grain, and Stover- Yield Related Traits in Pearl Millet (Pennisetum Glaucum (L.) R. Br.).” The ScientificWorld Journal 2014: 1–14. doi:10.1155/2014/562327.

Karkkainen, Katri, Veikko Koski, and Outi Savolainen. 1996. “Geographical Variation in the Inbreeding Depression of Scots Pine.” Evolution 50 (1): 111. doi:10.2307/2410785.

Karlsson, Sylvia B., and Indra K. Vasil. 1986. “Growth, Cytology and Flow Cytometry of Embryogenic Cell Suspension Cultures of Panicum Maximum Jacq. and Pennisetum Purpureum Schum.” Journal of Plant Physiology 123. Gustav Fischer Verlag, Stuttgart: 211–27. doi:10.1016/S0176-1617(86)80072-4.

Kearsey, Mike J., and A. G. L. Farquhar. 1998. “QTL Analysis in Plants ; Where Are We Now?” Heredity 80: 137–42.

Kellis, Manolis, Bruce W Birren, and Eric S Lander. 2004. “Proof and Evolutionary Analysis of Ancient Genome Duplication in the Yeast Saccharomyces Cerevisiae.” Nature 428 (6983): 617–24. doi:10.1038/nature02424.

Kent, W James. 2002. “BLAT — The BLAST -Like Alignment Tool.” Genome Research 12: 656–64. doi:10.1101/gr.229202.

191

Khan, Z. R., J. A. Pickett, L. Wadhams, and F. Muyekho. 2001. “Habitat Management Strategies for the Control of Cereal Stemborers and Striga in Maize in Kenya.” Insect Science and Its Application 21 (4): 375–80. doi:10.1017/S1742758400008481.

Kim, Changsoo, Hui Guo, Wenqian Kong, Rahul Chandnani, Lan-Shuan Shuang, and Andrew H. Paterson. 2015. “Application of Genotyping by Sequencing Technology to a Variety of Crop Breeding Programs.” Plant Science 242. Elsevier Ireland Ltd: 14–22. doi:10.1016/j.plantsci.2015.04.016.

Kojima, Shoko, Yuji Takahashi, Yasushi Kobayashi, Lisa Monna, Takuji Sasaki, Takashi Araki, and Masahiro Yano. 2002. “Hd3a, a Rice Ortholog of the Arabidopsis FT Gene, Promotes Transition to Flowering Downstream of Hd1 under Short-Day Conditions.” Plant & Cell Physiology 43 (10): 1096–1105. doi:10.1093/pcp/pcf156.

Krzywinski, Martin, Jacqueline Schein, Inanc Birol, Joseph Connors, Randy Gascoyne, Doug Horsman, Steven J Jones, and Marco A Marra. 2009. “Circos: An Information Aesthetic for Comparative Genomics.” Genome Research 19 (9). Cold Spring Harbor Lab: 1639–45.

Kuittinen, H, M J Sillanpa, and O Savolainen. 1997. “Genetic Basis of Adaptation : Flowering Time in Arabidopsis Thaliana.” Theoretical and Applied Genetics 95: 573–83.

Kumar, Lalitha Sunil. 1999. “DNA Markers in Plant Improvement: An Overview.” Biotechnology Advances 17 (September): 143–82.

Lagercrantz, Ulf, Joanna Putterill, George Coupland, and Derek Lydiate. 1996. “Comparative Mapping in Arabidopsis and Brassica, Fine Scale Genome Collinearity and Congruence of Genes Controlling Flowering Time.” The Plant Journal 9 (1): 13–20.

Lambert, Allison M., Abraham J. Miller-Rushing, and David W. Inouye. 2010. “Changes in Snowmelt Date and Summer Precipitation Affect the Flowering Phenology of Erythronium Grandiflorum (Glacier Lily; Liliaceae).” American Journal of Botany 97 (9): 1431–37. doi:10.3732/ajb.1000095.

Langmead, Ben, and Steven L Salzberg. 2012. “Fast Gapped-Read Alignment with Bowtie 2.” Nature Methods 9 (4): 357–59. doi:10.1038/nmeth.1923.

Laurie, David A., Nicky Pratchett, Jeremy H Bezant, and John W Snapet. 1994. “Genetic Analysis of a Photoperiod Response Gene on the Short Arm of Chromosome 2(2H) of Hordeum Vulgare (Barley).” Heredity 72: 619–27.

192

Lee, Ilha, Anthony Bleecker, and Richard Amasino. 1993. “Analysis of Naturally Occurring Late Flowering in Arabidopsis Thaliana.” Molecular and General Genetics MGG 237. Springer: 171–76.

Lewis, Michael W., Michelle E. Leslie, Emilee H. Fulcher, Lalitree Darnielle, Patrick Healy, Ji-Young Youn, and Sarah J. Liljegren. 2010. “The SERK1 Receptor-like Kinase Regulates Organ Separation in Arabidopsis Flowers.” Plant Journal 62 (5): 817–28. doi:10.1111/j.1365-313X.2010.04194.x.

Li, Heng, and Richard Durbin. 2010. “Fast and Accurate Long-Read Alignment with Burrows-Wheeler Transform.” Bioinformatics (Oxford, England) 26 (5). Wellcome Trust Sanger Institute, Wellcome Genome Campus, Cambridge, CB10 1SA, UK. DOI - 10.1093/bioinformatics/btp698 SRC - Pubmed ID2 - 20080505 FG - 0: 589–95. doi:10.1093/bioinformatics/btp698.

Li, Heng, Bob Handsaker, Alec Wysoker, Tim Fennell, Jue Ruan, Nils Homer, Gabor Marth, Goncalo Abecasis, Richard Durbin, and 1000 Genome Project Data Processing Subgroup. 2009. “The Sequence Alignment/map Format and SAMtools.” Bioinformatics 25 (16). Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Cambridge, CB10 1SA, UK, Broad Institute of MIT and Harvard, Cambridge, MA 02141, USA. DOI - 10.1093/bioinformatics/btp352 SRC - Pubmed ID2 - 19505943 FG - 0: 2078–79. doi:10.1093/bioinformatics/btp352.

Li, Weizhong, and Adam Godzik. 2006. “Cd-Hit: A Fast Program for Clustering and Comparing Large Sets of Protein or Nucleotide Sequences.” Bioinformatics 22 (13): 1658–59. doi:10.1093/bioinformatics/btl158.

Li, Wuxing, Yun Zhou, Xing Liu, Peng Yu, Jerry D Cohen, and Elliot M Meyerowitz. 2013. “LEAFY Controls Auxin Response Pathways in Floral Primordium Formation.” Science Signaling 6 (270): 1–6. doi:10.1126/scisignal.2003937.

Li, Yuan, Kai Yang, Wei Yang, Liwei Chu, Chunhai Chen, Bo Zhao, Yisong Li, et al. 2017. “Identification of QTL and Qualitative Trait Loci for Agronomic Traits Using SNP Markers in the Adzuki Bean.” Frontiers in Plant Science 8: 1–14. doi:10.3389/fpls.2017.00840.

Lindhout, Pim, Sjaak Van Heusden, Gerard Pet, Johan W. van Ooijen, Hans Sandbrink, Ruud Verkerk, Ria Vrielink, and Pim Zabel. 1994. “Perspectives of Molecular Marker Assisted Breeding for Earliness in Tomato.” Euphytica 79 (3). Springer: 279–86.

Loope, Lloyd L, Ole Hamann, and Charles P Stone. 1988. “Comparative Conservation Biology of Oceanic Archipelagoes: Hawaii and the Galapagos.” BioScience 38 (4). JSTOR: 272–82.

193

López, Yolanda, Aleksey Kurashev, Christine Chase, Maria Gallo, Lynn E. Sollenberger, Fredy Altpeter, and Jianping Wang. 2018. “Developing and Validating Microsatellite Markers in Elephant Grass (Pennisetum Purpureum S.).” Euphytica 214 (185): 1–`3. doi:10.1007/s10681-018-2256-6.

López, Yolanda, Jeffery Seib, Kenneth R. Woodard, Karen Chamusco, Lynn E. Sollenberger, Maria Gallo, S. Luke Flory, and Christine Chase. 2014. “Genetic Diversity of Biofuel and Naturalized Napiergrass (Pennisetum Purpureum).” Invasive Plant Science and Management 7 (2). Weed Science Society of America 810 East 10th Street, Lawrence, KS 66044-8897: 229–36. doi:10.1614/IPSM-D-13-00085.1.

Lowe, Andrew J, W Thorpe, A Teale, and J Hanson. 2003. “Characterisation of Germplasm Accessions of Napier Grass (Pennisetum Purpureum and P. Purpureum× P. Glaucum Hybrids) and Comparison with Farm Clones Using RAPD.” Genetic Resources and Crop Evolution 50 (2). Springer: 121–32.

Lu, Fei, Alexander E. Lipka, Jeffrey C. Glaubitz, Robert J. Elshire, Jerome H. Cherney, Michael D. Casler, Edward S. Buckler, and Denise E. Costich. 2013. “Switchgrass Genomic Diversity, Ploidy, and Evolution: Novel Insights from a Network-Based SNP Discovery Protocol.” PLoS Genetics 9 (1): 1–14. doi:10.1371/journal.pgen.1003215.

Lu, Hongfeng, Tao Lin, Joël Klein, Shenhao Wang, Jianjian Qi, Qian Zhou, Jinjing Sun, Zhonghua Zhang, Yiqun Weng, and Sanwen Huang. 2014. “QTL-Seq Identifies an Early Flowering QTL Located near Flowering Locus T in Cucumber.” Theoretical and Applied Genetics 127. Springer: 1–9. doi:10.1007/s00122-014- 2313-z.

Ma, Xue Feng, Elaine Jensen, Nickolai Alexandrov, Maxim Troukhan, Liping Zhang, Sian Thomas-Jones, Kerrie Farrar, et al. 2012. “High Resolution Genetic Mapping by Genome Sequencing Reveals Genome Duplication and Tetraploid Genetic Structure of the Diploid Miscanthus Sinensis.” PLoS ONE 7 (3): 1–11. doi:10.1371/journal.pone.0033821.

Maheswaran, M, Ning Huang, S R Sreerangasamy, and Susan R McCouch. 2000. “Mapping Quantitative Trait Loci Associated with Days to Flowering and Photoperiod Sensitivity in Rice (Oryza Sativa L.).” Molecular Breeding 6. Springer: 145–55.

Makumburage, G. Buddhika, and Ann E. Stapleton. 2011. “Phenotype Uniformity in Combined-Stress Environments Has a Different Genetic Architecture than in Single-Stress Treatments.” Frontiers in Plant Science 2: 1–8. doi:10.3389/fpls.2011.00012.

194

Martel, E., D De Nay, S Siljak-Yakoviev, S. Brown, A. Sarr, D. De Nay, S. Siljak- Yakovlev, et al. 1997. “Genome Size Variation and Basic Chromosome Number in Pearl Millet and Fourteen Related Pennisetum Species.” Journal of Heredity 88. Am Genetic Assoc: 139–43.

Martel, E., V. Poncet, F. Lamy, S. Siljak-Yakovlev, B. Lejeune, and A. Sarr. 2004. “Chromosome Evolution of Pennisetum Species (Poaceae): Implications of ITS Phylogeny.” Plant Systematics and Evolution 249: 139–49. doi:10.1007/s00606- 004-0191-6.

Martel, E., A Ricroch, and A Sarr. 1996. “Assessment of Genome Organization among Diploid Species (2n = 2x = 14) Belonging to Primary and Tertiary Gene Pools of Pearl Millet Using Fluorescent in Situ Hybridization with rDNA Probes.” Genome 39. NRC Research Press: 680–87.

Martin, K. L., P. J. Hodgen, K. W. Freeman, Ricardo Melchiori, D. B. Arnall, R. K. Teal, R. W. Mullen, et al. 2005. “Plant-to-Plant Variability in Corn Production.” Agronomy Journal 97 (6): 1603–11. doi:10.2134/agronj2005.0129.

Mathew, Lisa S, Manuel Spannagl, Ameena Al-malki, Binu George, Maria F Torres, Eman K Al-dous, Eman K Al-azwani, et al. 2014. “A First Genetic Map of Date Palm (Phoenix Dactylifera) Reveals Long-Range Genome Structure Conservation in the Palms.” BMC Genomics 15 (285): 1–10.

Matsubara, Kazuki, I. Kono, K. Hori, Y. Nonoue, N. Ono, A. Shomura, T. Mizubayashi, et al. 2008. “Novel QTLs for Photoperiodic Flowering Revealed by Using Reciprocal Backcross Inbred Lines from Crosses between Japonica Rice Cultivars.” Theoretical and Applied Genetics 117: 935–45. doi:10.1007/s00122- 008-0833-0.

Matsubara, Kazuki, Eri Ogiso-Tanaka, Kiyosumi Hori, Kaworu Ebana, Tsuyu Ando, and Masahiro Yano. 2012. “Natural Variation in Hd17, a Homolog of Arabidopsis ELF3 That Is Involved in Rice Photoperiodic Flowering.” Plant and Cell Physiology 53 (4): 709–16. doi:10.1093/pcp/pcs028.

McCouch, Susan R, Y. G. Cho, M. Yano, E. Paul, M. Blinstrub, H. Morishima, and T. Kinoshita. 1997. “Report on QTL Nomenclature.” Rice Genetics Newsletter 14: 11–13.

Mckendry, Peter. 2002. “Energy Production from Biomass (Part 1): Overview of Biomass.” Bioresource Technology 83: 37–46.

McKenna, Aaron, Matthew Hanna, Eric Banks, Andrey Sivachenko, Kristian Cibulskis, Andrew Kernytsky, Kiran Garimella, et al. 2010. “The Genome Analysis Toolkit: A MapReduce Framework for Analyzing next-Generation DNA Sequencing Data.” Genome Research 20 (9): 1297–1303. doi:10.1101/gr.107524.110.

195

Melo, Arthur T. O., Radhika Bartaula, and Iago Hale. 2016. “GBS-SNP-CROP: A Reference-Optional Pipeline for SNP Discovery and Plant Germplasm Characterization Using Variable Length, Paired-End Genotyping-by-Sequencing Data.” BMC Bioinformatics 17 (29): 1–15. doi:10.1186/s12859-016-0879-y.

Miles, Cecelia M., and Marta Wayne. 2008. “Quantitative Trait Locus ( QTL ) Analysis.” Nature Education 1 (208): 1–7.

Morris, Karl, Sarah Thornber, Lesley Codrai, Christine Richardson, Adam Craig, Ari Sadanandom, Brian Thomas, and Stephen Jackson. 2010. “DAY NEUTRAL FLOWERING Represses CONSTANS to Prevent Arabidopsis Flowering Early in Short Days.” The Plant Cell 22: 1118–28. doi:10.1105/tpc.109.066605.

Muldoon, D K, and C J Pearson. 1979. “The Hybrid between Pennisetum Americanum and Pennisetum Purpureum.” Herbage Abstracts 49 (5): 189–99.

Murray, Seth C., William L. Rooney, Sharon E. Mitchell, Arun Sharma, Patricia E. Klein, John E. Mullet, and Stephen Kresovich. 2008. “Genetic Improvement of Sorghum as a Biofuel Feedstock: II. QTL for Stem and Leaf Structural Carbohydrates.” Crop Science 48: 2180–93.

Nagy, Istvan, Susanne Barth, Jeanne Mehenni-Ciz, Michael T Abberton, and Dan Milbourne. 2013. “A Hybrid next Generation Transcript Sequencing-Based Approach to Identify Allelic and Homeolog-Specific Single Nucleotide Polymorphisms in Allotetraploid White Clover.” BMC Genomics 14 (100): 1–19. doi:10.1186/1471-2164-14-100.

Negawo, Alemayehu Teressa, Abel Teshome, Alok Kumar, Jean Hanson, and Chris S Jones. 2017. “Opportunities for Napier Grass (Pennisetum Purpureum) Improvement Using Molecular Genetics.” Agronomy 7 (28): 1–21.

Noble Research Institute. 2007. “Switchgrass Establishment Requires Patience.” Ag News & Views. https://www.noble.org/news/publications/ag-news-and- views/2007/july/switchgrass-establishment-requires-patience/.

Obok, Ekemini Edet, Micheal E Aken Ova, and Godfrey A Iwo. 2012. “Forage Potentials of Interspecific Hybrids between Elephant Grass Selections and Cultivated Pearl Millet Genotypes of Nigerian Origin.” Journal of Plant Breeding and Crop Science 4 (9): 136–43. doi:10.5897/JPBCS12.019.

Osborn, Thomas C, C. Kole, I. A. P. Parkin, A. G. Sharpe, M. Kuiper, D. J. Lydiate, and Martin Trick. 1997. “Comparison of Flowering Time Genes in Brassica Rapa, B. Napus and Arabidopsis Thaliana.” Genetics 146: 1123–29.

Osgood, RV V, Wayne W. Hanna, and TL L Tew. 1997. “Hybrid Seed Production of Pearl Millet × Napiergrass Triploid Hybrids.” Crop Science 37. Crop Science Society of America: 998–99.

196

Paillard, S., T. Schnurbusch, M. Winzeler, M. Messmer, P. Sourdille, O. Abderhalden, B. Keller, and G. Schachermayr. 2003. “An Integrative Genetic Linkage Map of Winter Wheat (Triticum Aestivum L.).” Theoretical and Applied Genetics 107: 1235–42. doi:10.1007/s00122-003-1361-6.

Pandey, Manish K., Emmanuel Monyo, Peggy Ozias-Akins, Xuanquiang Liang, Patricia Guimares, Shyam N. Nigam, Hari D. Upadhyaya, et al. 2012. “Advances in Arachis Genomics for Peanut Improvement.” Biotechnology Advances 30. Elsevier Inc.: 639–51. doi:10.1016/j.biotechadv.2011.11.001.

Paterson, Andrew H., John E Bowers, Rémy Bruggmann, Inna Dubchak, Jane Grimwood, Heidrun Gundlach, Georg Haberer, et al. 2009. “The Sorghum Bicolor Genome and the Diversification of Grasses.” Nature 457 (7229): 551–56. doi:10.1038/nature07723.

Paudel, Dev R., Baskaran Kannan, Xiping Yang, Karen R. Harris-Shultz, Mahendar Thudi, Rajeev K. Varshney, Fredy Altpeter, and Jianping Wang. 2018. “Surveying the Genome and Constructing a High-Density Genetic Map of Napiergrass (Cenchrus Purpureus Schumach).” Scientific Reports 8 (1). Springer US: 1–11. doi:10.1038/s41598-018-32674-x.

Paux, Etienne, Pierre Sourdille, Ian Mackay, and Catherine Feuillet. 2012. “Sequence- Based Marker Development in Wheat: Advances and Applications to Breeding.” Biotechnology Advances 30. Elsevier Inc.: 1071–88. doi:10.1016/j.biotechadv.2011.09.015.

Peng, Ze, Wen Fan, Liping Wang, Dev R. Paudel, Dante Leventini, Barry L. Tillman, and Jianping Wang. 2017. “Target Enrichment Sequencing in Cultivated Peanut (Arachis Hypogaea L.) Using Probes Designed from Transcript Sequences.” Molecular Genetics and Genomics. Springer Berlin Heidelberg, 1–11. doi:10.1007/s00438-017-1327-z.

Phan, Huyen T.T., Kasia Rybak, Stefania Bertazzoni, Eiko Furuki, Eric Dinglasan, Lee T. Hickey, Richard P. Oliver, and Kar Chun Tan. 2018. “Novel Sources of Resistance to Septoria Nodorum Blotch in the Vavilov Wheat Collection Identified by Genome-Wide Association Studies.” Theoretical and Applied Genetics 131. Springer Berlin Heidelberg: 1223–38. doi:10.1007/s00122-018-3073-y.

Phukan, Ujjal J., Gajendra S. Jeena, and Rakesh K. Shukla. 2016. “WRKY Transcription Factors: Molecular Regulation and Stress Responses in Plants.” Frontiers in Plant Science 7: 1–14. doi:10.3389/fpls.2016.00760.

Poland, Jesse A., Patrick J Brown, Mark E Sorrells, and Jean-Luc Jannink. 2012. “Development of High-Density Genetic Maps for Barley and Wheat Using a Novel Two-Enzyme Genotyping-by-Sequencing Approach.” PloS One 7 (2). Public Library of Science: 1–8.

197

Poland, Jesse A., J Endelman, J Dawson, J Rutkoski, S Y Wu, Y Manes, S Dreisigacker, et al. 2012. “Genomic Selection in Wheat Breeding Using Genotyping-by-Sequencing.” Plant Genome 5 (3): 103–13. doi:Doi 10.3835/Plantgenome2012.06.0006.

Poland, Jesse A., and Trevor W Rife. 2012. “Genotyping-by-Sequencing for Plant Breeding and Genetics.” The Plant Genome 5 (3). Crop Science Society of America: 92–102.

Pootakham, Wirulda, Nukoon Jomchai, Panthita Ruang-areerate, Jeremy R. Shearman, Chutima Sonthirod, Duangjai Sangsrakru, Somvong Tragoonrung, and Sithichoke Tangphatsornruang. 2015. “Genome-Wide SNP Discovery and Identification of QTL Associated with Agronomic Traits in Oil Palm Using Genotyping-by-Sequencing (GBS).” Genomics 105. Elsevier B.V.: 288–95. doi:10.1016/j.ygeno.2015.02.002.

Powell, Jerrel B, and Glenn W. Burton. 1966. “A Suggested Commercial Method of Producing an Interspecific Hybrid Forage in Pennisetum.” Crop Science 6. Crop Science Society of America: 378–79.

Powell, Wayne, Gordon C Machray, and Jim Provan. 1996. “Polymorphism Revealed by Simple Sequence Repeats.” Trends in Plant Science 1 (7). Elsevier: 215–22.

Price, David L., and Michael D. Casler. 2014. “Divergent Selection for Secondary Traits in Upland Tetraploid Switchgrass and Effects on Sward Biomass Yield.” Bioenergy Research 7: 329–37. doi:10.1007/s12155-013-9374-8.

Pruitt, Kim D., Tatiana Tatusova, and Donna R. Maglott. 2007. “NCBI Reference Sequences (RefSeq): A Curated Non-Redundant Sequence Database of Genomes, Transcripts and Proteins.” Nucleic Acids Research 33: 501–4. doi:10.1093/nar/gkl842.

Punnuri, Somashekhar M., Jason G. Wallace, Joseph E. Knoll, Katie E. Hyma, Sharon E. Mitchell, Edward S. Buckler, Rajeev K. Varshney, Bharat P. Singh, and T H E Plant Genome. 2016. “Development of a High-Density Linkage Map and Tagging Leaf Spot Resistance in Pearl Millet Using GBS Markers.” Plant Genome 9 (2): 1–13. doi:10.3835/plantgenome2015.10.0106.

Ra, Kaien, Fumitaka Shiotsu, Jun Abe, and Shigenori Morita. 2012. “Biomass Yield and Nitrogen Use Efficiency of Cellulosic Energy Crops for Ethanol Production.” Biomass and Bioenergy 37. Elsevier: 330–34.

Rafalski, Antoni. 2002. “Applications of Single Nucleotide Polymorphisms in Crop Genetics.” Current Opinion in Plant Biology 5. Elsevier: 94–100.

198

Rajaram, Vengaldas, Thirunavukkarasu Nepolean, Senapathy Senthilvel, Rajeev K. Varshney, Vincent Vadez, Rakesh K Srivastava, Trushar M Shah, Ambawat Supriya, Sushil Kumar, and Basava Ramana Kumari. 2013. “Pearl Millet [Pennisetum Glaucum (L.) R. Br.] Consensus Linkage Map Constructed Using Four RIL Mapping Populations and Newly Developed EST-SSRs.” BMC Genomics 14 (159). BioMed Central Ltd: 1–15.

Ram, Hari Har, Hari Har Ram, and Rakesh Yadava. 2007. Genetic Resources and Seed Enterprises: Management and Policies. New India Publishing.

Reis, Gabriela Barreto dos, Amanda Teixeira Mesquita, Giovana Augusta Torres, Larissa Fonseca Andrade-Vieira, Antônio Vander Pereira, and Lisete Chamma Davide. 2014. “Genomic Homeology between Pennisetum Purpureum and Pennisetum Glaucum (Poaceae).” Comparative Cytogenetics 8 (3): 199–209. doi:10.3897/CompCytogen.v8i3.7732.

Rice, Peter, Lan Longden, and Alan Bleasby. 2000. “EMBOSS: The European Molecular Biology Open Software Suite.” Trends in Genetics 16 (6): 276–77. doi:10.1016/S0168-9525(00)02024-2.

Richmond, Todd A., and Anthony B. Bleecker. 1999. “A Defect in Beta-Oxidation Causes Abnormal Inflorescence Development in Arabidopsis.” The Plant Cell 11: 1911–24. doi:10.1105/tpc.11.10.1911.

Rosyara, Umesh R., Walter S. De Jong, David S. Douches, and Jeffrey B. Endelman. 2016. “Software for Genome-Wide Association Studies in Autopolyploids and Its Application to Potato.” The Plant Genome 9 (2): 1–10. doi:10.3835/plantgenome2015.08.0073.

Rozen, S, and H Skaletsky. 2000. “Primer3 on the WWW for General Users and for Biologist Programmers.” Methods in Molecular Biology 132. Whitehead Institute for Biomedical Research, Cambridge, MA, USA. SRC - Pubmed ID2 - 10547847 FG - 0: 365–86.

Russell, Joanne, Martin Mascher, Ian K. Dawson, Stylianos Kyriakidis, Cristiane Calixto, Fabian Freund, Micha Bayer, et al. 2016. “Exome Sequencing of Geographically Diverse Barley Landraces and Wild Relatives Gives Insights into Environmental Adaptation.” Nature Genetics 48 (9). Nature Publishing Group: 1024–30. doi:10.1038/ng.3612.

Samach, Alon. 2012. “Control of Flowering.” In Plant Biotechnology and Agriculture:, First Edit, 387–404. Elsevier Inc. doi:10.1016/B978-0-12-381466-1.00025-0.

Schlegel, Rolf H J. 2009. Dictionary of Plant Breeding. CRC Press.

Schnable, Patrick S., Doreen Ware, RS Fulton, and JC Stein. 2009. “The B73 Maize Genome: Complexity, Diversity, and Dynamics.” Science 326 (5956): 1112–15.

199

Schneider, Katharina. 2005. “Mapping Populations and Principles of Genetic Mapping.” In The Handbook of Plant Genome Mapping: Genetic and Physical Mapping, edited by Khalid Meksem and Gunter Kahl, 1–21. WILEY-VCH Verlag GmbH & Co. KGaA. doi:10.1002/3527603514.ch1.

Schofield, Eileen K. 1989. “Effects of Introduced Plants and Animals on Island Vegetation: Examples from Galápagos Archipelago.” Conservation Biology 3 (3). Wiley Online Library: 227–39.

Semagn, Kassa, Raman Babu, Sarah Hearne, and Michael Olsen. 2014. “Single Nucleotide Polymorphism Genotyping Using Kompetitive Allele Specific PCR (KASP): Overview of the Technology and Its Application in Crop Improvement.” Molecular Breeding. doi:10.1007/s11032-013-9917-x.

Shen, Jiawei, Zhiqiang Li, Jianhua Chen, Zhijian Song, Zhaowei Zhou, and Yongyong Shi. 2016. “SHEsisPlus, a Toolset for Genetic Studies on Polyploid Species.” Scientific Reports 6. Nature Publishing Group: 1–10. doi:10.1038/srep24095.

Shepherd, M., M. Cross, M. J. Dieters, and R. Henry. 2003. “Genetic Maps for Pinus Elliottii Var. Elliottii and P. Caribaea Var. Hondurensis Using AFLP and Microsatellite Markers.” Theoretical and Applied Genetics 106: 1409–19. doi:10.1007/s00122-002-1185-9.

Simpson, Jared T, Kim Wong, Shaun D Jackman, Jacqueline E Schein, Steven J M Jones, Inanç Birol, Jared T Simpson, et al. 2009. “ABySS: A Parallel Assembler for Short Read Sequence Data.” Genome Research 19 (June): 1117–23. doi:10.1101/gr.089532.108.

Sinche, Marco. 2013. “Genetic Improvement of Elephantgrass (Pennisetum Purpureum Schum.) as a Bioenergy Feedstock.” University of Florida.

Sinche, Marco, Baskaran Kannan, Carlos Corsato, Dev R. Paudel, Jianping Wang, and Fredy Altpeter. 2018. “Development and Characterization of Napiergrass (Pennisetum Purpureum) Mapping Population for Flowering Time and Biomass Related Traits.” Submitted for Publication.

Singer, S. D., N. T. Krogan, and N. W. Ashton. 2007. “Clues about the Ancestral Roles of Plant MADS-Box Genes from a Functional Analysis of Moss Homologues.” Plant Cell Reports 26: 1155–69. doi:10.1007/s00299-007-0312-0.

Singh, Bharat P., Hari P Singh, and Eric Obeng. 2013. “Elephantgrass.” In Biofuel Crops: Production, Physiology and Genetics, edited by Bharat P Singh, 271–91. CAB International.

Smith, Rex L., and M K U Chowdhury. 1983. “Mitochondrial DNA Polymorphism in Male-Sterile and Fertile Cytoplasms of Pearl Millet.” Crop Science 29: 809–14.

200

Smith, Rex L., M. E. Schweder, M. K. U. Chowdhury, J. C. Seib, and S. C. Schank. 1993. “Development and Application of RFLP and RAPD DNA Markers in Genetic Improvement of Pennisetum for Biomass and Forage Production.” Biomass and Bioenergy 5 (1): 51–62.

Sollenberger, Lynn E., G M Prine, W R Ocumpaugh, Wayne W. Hanna, C S Jones, S C Schank, and R S Kalmbacher. 1989. “Registration of ‘Mott’dwarf Elephantgrass.” Crop Science 29. Crop Science Society of America: 827–28.

Sollenberger, Lynn E., Kenneth R. Woodard, João M.B. B. Vendramini, John E. Erickson, Kenneth A. Langeland, M. Kimberly Mullenix, Chaein Na, et al. 2014. “Invasive Populations of Elephantgrass Differ in Morphological and Growth Characteristics from Clones Selected for Biomass Production.” Bioenergy Research 7 (May): 1382–91. doi:10.1007/s12155-014-9478-9.

Somerville, Chris, Heather Youngs, Caroline Taylor, Sarah C Davis, and Stephen P Long. 2010. “Feedstocks for Lignocellulosic Biofuels.” Science 329: 790–93.

Song, Xiaoming, Tingting Ge, Ying Li, and Xilin Hou. 2015. “Genome-Wide Identification of SSR and SNP Markers from the Non-Heading Chinese Cabbage for Comparative Genomic Analyses.” BMC Genomics 16 (328): 1–18. doi:10.1186/s12864-015-1534-0.

Stinchcombe, JR, and HE Hoekstra. 2008. “Combining Population Genomics and Quantitative Genetics : Finding the Genes Underlying Ecologically Important Traits.” Heredity 100: 158–70. doi:10.1038/sj.hdy.6800937.

Stone, Eric A., and Julien F. Ayroles. 2009. “Modulated Modularity Clustering as an Exploratory Tool for Functional Genomic Inference.” PLoS Genetics 5 (5): 1–13. doi:10.1371/journal.pgen.1000479.

Struwig, M. 2007. “Fingerprinting Pennisetum Purpureum Schumach. Varieties and Cultivars Using AFLP Analyses.” North West University.

Suárez-López, P, K Wheatley, F Robson, H Onouchi, F Valverde, and G Coupland. 2001. “CONSTANS Mediates between the Circadian Clock and the Control of Flowering in Arabidopsis.” Nature 410: 1116–20. doi:10.1038/35074138.

Tamura, Koichiro, Glen Stecher, Daniel Peterson, Alan Filipski, and Sudhir Kumar. 2013. “MEGA6: Molecular Evolutionary Genetics Analysis Version 6.0.” Molecular Biology and Evolution, 2725–29. doi:10.1093/molbev/mst197.

Taylor, D R, and P K Ingvarsson. 2003. “Common Features of Segregation Distortion in Plants and Animals.” Genetica 117: 27–35. doi:10.1023/A:1022308414864.

201

Taylor, M G, and I K Vasil. 1987. “Analysis of DNA Size , Content and Cell Cycle in Leaves of Napier Grass (Pennisetum Purpureum Schum .).” Theoretical and Applied Genetics 74: 681–86.

Thiel, T., W. Michalek, Rajeev K. Varshney, and A. Graner. 2003. “Exploiting EST Databases for the Development and Characterization of Gene-Derived SSR- Markers in Barley (Hordeum Vulgare L.).” Theoretical and Applied Genetics 106. Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstrasse 3, D- 06466 Gatersleben, Germany. DOI - 10.1007/s00122-002-1031-0 SRC - Pubmed ID2 - 12589540 FG - 0: 411–22. doi:10.1007/s00122-002-1031-0.

Thompson, J. B. 1919. Napier and Merker Grass : Two New Forage Crops for Florida. Gainesville: University of Florida, Agricultural Experiment Station.

Tollenaar, M, and J Wu. 1999. “Yield Improvement in Temperate Maize Is Attributed to Greater Stress Tolerance.” Crop Science 39: 1597–1604. doi:10.2135/cropsci1999.3961597x.

Tong, Xinggang, Laurence H Smith, and Perry L Mccarty. 1990. “Methane Fermentation of Selected Lignocellulosic Materials.” Biomass 21: 239–55.

Trick, Martin, Yan Long, Jinling Meng, and Ian Bancroft. 2009. “Single Nucleotide Polymorphism (SNP) Discovery in the Polyploid Brassica Napus Using Solexa Transcriptome Sequencing.” Plant Biotechnology Journal 7: 334–46. doi:10.1111/j.1467-7652.2008.00396.x.

Tsai, Yu Chang, Nikkí A. Delk, Naweed I. Chowdhury, and Janet Braam. 2007. “Arabidopsis Potential Calcium Sensors Regulate Nitric Oxide Levels and the Transition to Flowering.” Plant Signaling and Behavior 2 (6): 446–54. doi:10.4161/psb.2.6.4695.

Uitdewilligen, Jan G A M L, Anne Marie A Wolters, Bjorn B. D’hoop, Theo J A Borm, Richard G F Visser, and Herman J. van Eck. 2013. “A next-Generation Sequencing Method for Genotyping-by-Sequencing of Highly Heterozygous Autotetraploid Potato.” PLoS ONE 8 (5): 1–14. doi:10.1371/journal.pone.0062355.

USDA, ARS, and NGRP. 1954. “PI 215603.” Germplasm Resources Information Network - (GRIN). [Online Database] National Germplasm Resources Laboratory, Beltsville, Maryland. http://www.ars-grin.gov/cgi- bin/npgs/acc/display.pl?1176422.

———. 1963. “PI 288787.” Germplasm Resources Information Network - (GRIN). [Online Database] National Germplasm Resources Laboratory, Beltsville, Maryland. http://www.ars-grin.gov/cgi-bin/npgs/acc/display.pl?1218439.

202

van Ooijen, Johan W. 1992. “Accuracy of Mapping Quantitative Trait Loci in Autogamous Species.” Theoretical and Applied Genetics 84: 803–11. doi:10.1007/BF00227388.

———. 2006. “JoinMap 4.0: Software for the Calculation of Genetic Linkage Maps in Experimental Populations.” Wageningen, Netherlands: Kyazma BV.

Varshney, Rajeev K., Chengcheng Shi, Mahendar Thudi, Cedric Mariac, Jason Wallace, Peng Qi, He Zhang, et al. 2017. “Pearl Millet Genome Sequence Provides a Resource to Improve Agronomic Traits in Arid Environments.” Nature Biotechnology 35 (10): 969–76.

Veeckman, E, K Vandepoele, T Asp, and T Ruttink. 2016. “Genomic Variation in the FT Gene Family of Perennial Ryegrass (Lolium Perenne).” In Breeding in a World of Scarcity: Proceedings of the 2015 Meeting of the Section “Forage Crops and Amenity Grasses” of Eucarpia, 121–26. doi:10.1007/978-3-319-28932-8.

Vicente-Chandler, Jose, Servando Silva, and Jacinto Figarella. 1959. “The Effect of Nitrogen Fertilization and Frequency of Cutting on the Yield and Composition of Three Tropical Grasses.” Agronomy Journal 51. American Society of Agronomy: 202–6.

Voorrips, Roeland E. 2002. “MapChart: Software for the Graphical Presentation of Linkage Maps and QTLs.” The Journal of Heredity, 77–78. doi:10.1093/jhered/93.1.77.

Wang, Cuiling, Yanhui Chen, Lixia Ku, Tiegu Wang, Zhaohui Sun, Fangfang Cheng, and Liancheng Wu. 2010. “Mapping QTL Associated with Photoperiod Sensitivity and Assessing the Importance of QTL×environment Interaction for Flowering Time in Maize.” PLoS ONE 5 (11): 8–10. doi:10.1371/journal.pone.0014068.

Wang, Jia, Zhitong Chen, Surong Jin, Zhongli Hu, Yibing Huang, and Ying Diao. 2017. “Development and Characterization of Simple Sequence Repeat (SSR) Markers Based on a Full-Length cDNA Library of Napier Grass (Pennisetum Purpureum Schum).” Genes and Genomics 39 (12). The Genetics Society of Korea: 1297– 1305. doi:10.1007/s13258-017-0536-5.

Wang, Shengchu, CJ Basten, P Gaffney, and Z-B Zeng. 2005. “Windows QTL Cartographer 2.5. User Manual.” International Immunology. Department of Statistics, North Carolina State University, Raleigh, NC. doi:10.1093/intimm/dxs150.

Wang, Yan, Silong Sun, Bo Liu, Hui Wang, Jie Deng, Yongcui Liao, Qian Wang, Feng Cheng, Xiaowu Wang, and Jian Wu. 2011. “A Sequence-Based Genetic Linkage Map as a Reference for Brassica Rapa Pseudochromosome Assembly.” BMC Genomics 12 (1): 239. doi:10.1186/1471-2164-12-239.

203

Wanjala, Bramwel W., Meshack Obonyo, Francis N. Wachira, Alice Muchugi, Margaret Mulaa, Jagger Harvey, Robert A. Skilton, Janice Proud, and Jean Hanson. 2013. “Genetic Diversity in Napier Grass (Pennisetum Purpureum) Cultivars: Implications for Breeding and Conservation.” AoB PLANTS 5 (January): 1–10. doi:10.1093/aobpla/plt022.

Woodard, Kenneth R., and Lynn E. Sollenberger. 2012. “Production of Biofuel Crops in Florida: Elephantgrass.” SS-AGR-297.

Wu, Changyin, Changjun You, Caishun Li, Tuan Long, Guoxing Chen, Mary E Byrne, and Qifa Zhang. 2008. “RID1, Encoding a Cys2/His2-Type Zinc Finger Transcription Factor, Acts as a Master Switch from Vegetative to Floral Development in Rice.” Proceedings of the National Academy of Sciences of the United States of America 105 (35): 12915–20.

Wu, K. K., W. Burnquist, M. E. Sorrells, T. L. Tew, P. H. Moore, and S. D. Tanksley. 1992. “The Detection and Estimation of Linkage in Polyploids Using Single-Dose Restriction Fragments.” Theoretical and Applied Genetics 83: 294–300. doi:10.1007/BF00224274.

Wu, S, J Yang, Y Huang, Y Li, T Yin, S D Wullschleger, G a Tuskan, and R Wu. 2010. “An Improved Approach for Mapping Quantitative Trait Loci in a Pseudo- Testcross: Revisiting a Poplar Mapping Study.” Bioinform Biol Insights 4: 1–8. doi:10.4137/BBI.S13402.TYPE.

Xiao-ping, Lu, Yun Jin-feng, Gao Cui-ping, and Surya Acharya. 2011. “Quantitative Trait Loci Analysis of Economically Important Traits in Sorghum Bicolor × S. Sudanense Hybrid.” Canadian Journal of Plant Science 91: 81–90. doi:10.4141/cjps09112.

Xie, S. Q., J. Y. Feng, and Y. M. Zhang. 2014. “Linkage Group Correction Using Epistatic Distorted Markers in F2 and Backcross Populations.” Heredity 112: 479–88. doi:10.1038/hdy.2013.127.

Xie, Wengang, Xinquan Zhang, Hongwei Cai, Linkai Huang, Yan Peng, and Xiao Ma. 2011. “Genetic Maps of SSR and SRAP Markers in Diploid Orchardgrass (Dactylis Glomerata L.) Using the Pseudo-Testcross Strategy.” Genome 54: 212– 21. doi:10.1139/G10-111.

Xie, Xin-Ming, Feng Zhou, Xiang-Qian Zhang, and Ju-Ming Zhang. 2009. “Genetic Variability and Relationship between MT-1 Elephant Grass and Closely Related Cultivars Assessed by SRAP Markers.” Journal of Genetics 88 (3). Springer: 281–90.

Xu, Shizhong. 2008. “Quantitative Trait Locus Mapping Can Benefit from Segregation Distortion.” Genetics 180: 2201–8. doi:10.1534/genetics.108.090688.

204

Xu, Yang, Pengcheng Li, Zefeng Yang, and Chenwu Xu. 2017. “Genetic Mapping of Quantitative Trait Loci in Crops.” Crop Journal 5. Elsevier B.V.: 175–84. doi:10.1016/j.cj.2016.06.003.

Yadav, R. S., F. R. Bidinger, C. T. Hash, Y. P. Yadav, O. P. Yadav, S. K. Bhatnagar, and C. J. Howarth. 2003. “Mapping and Characterisation of QTL X E Interactions for Traits Determining Grain and Stover Yield in Pearl Millet.” Theoretical and Applied Genetics 106: 512–20. doi:10.1007/s00122-002-1081-3.

Yang, Xiping, Md. S. Islam, Sushma Sood, Stephanie Maya, Erik A. Hanson, Jack Comstock, and Jianping Wang. 2018. “Identifying Quantitative Trait Loci (QTLs) and Developing Diagnostic Markers Linked to Orange Rust Resistance in Sugarcane (Saccharum Spp.).” Frontiers in Plant Science 9 (350): 1–10. doi:10.3389/fpls.2018.00350.

Yang, Xiping, Jian Song, James Todd, Ze Peng, Dev R. Paudel, Ziliang Luo, Xiaokai Ma, et al. 2018. “Target Enrichment Sequencing of 307 Germplasm Accessions Identified Ancestry of Ancient and Modern Hybrids and Signatures of Adaptation and Selection in Sugarcane ( Saccharum Spp.), a ‘sweet’ Crop with ‘bitter’ Genomes.” Plant Biotechnology Journal, 1–11. doi:10.1111/pbi.12992.

Yang, Xiping, Jian Song, Qian You, Dev R. Paudel, Jisen Zhang, and Jianping Wang. 2017. “Mining Sequence Variations in Representative Polyploid Sugarcane Germplasm Accessions.” BMC Genomics 18 (594). BMC Genomics: 1–16. doi:10.1186/s12864-017-3980-3.

Yang, Xiping, Sushma Sood, Neil Glynn, Md Sariful Islam, Jack Comstock, and Jianping Wang. 2017. “Constructing High-Density Genetic Maps for Polyploid Sugarcane (Saccharum Spp.) and Identifying Quantitative Trait Loci Controlling Brown Rust Resistance.” Molecular Breeding 37 (116). Molecular Breeding: 1–12.

Yano, Masahiro, Yuichi Katayose, Motoyuki Ashikari, Utako Yamanouchi, Lisa Monna, Takuichi Fuse, Tomoya Baba, Kimiko Yamamoto, Yosuke Umehara, and Yoshiaki Nagamura. 2000. “Hd1, a Major Photoperiod Sensitivity Quantitative Trait Locus in Rice, Is Closely Related to the Arabidopsis Flowering Time Gene CONSTANS.” The Plant Cell Online 12. Am Soc Plant Biol: 2473–83.

Yano, Masahiro, Shoko Kojima, Yuji Takahashi, Hongxuan Lin, and Takuji Sasaki. 2001. “Genetic Control of Flowering Time in Rice, a Short-Day Plant.” Plant Physiology 127. Am Soc Plant Biol: 1425–29. doi:10.1104/pp.010710.1.

Yanofsky, Martin F., Hong Ma, John L. Bowman, Gary N. Drews, Kenneth A. Feldmann, and Elliot M. Meyerowitz. 1990. “The Protein Encoded by the Arabidopsis Homeotic Gene Agamous Resembles Transcription Factors.” Nature 346: 35–39. doi:10.1038/346035a0.

205

Yoshida, H., K. Hirano, T. Sato, N. Mitsuda, M. Nomoto, K. Maeo, E. Koketsu, et al. 2014. “DELLA Protein Functions as a Transcriptional Activator through the DNA Binding of the INDETERMINATE DOMAIN Family Proteins.” Proceedings of the National Academy of Sciences 111 (21): 7861–66. doi:10.1073/pnas.1321669111.

Yu, Jun, Songnian Hu, Jing Juan Jian Jun Wang, Gane Ka-shu Wong, Songgang Li, Bin Liu, Yajun Deng, et al. 2002. “A Draft Sequence of the Rice Genome (Oryza Sativa L. Ssp. Indica).” Science (New York, N.Y.) 296 (5565): 79–92. doi:10.1126/science.1068037.

Yu, Xiaoqing, and Shuying Sun. 2013. “Comparing a Few SNP Calling Algorithms Using Low-Coverage Sequencing Data.” BMC Bioinformatics 14 (274): 1–15. doi:10.1186/1471-2105-14-274.

Zahn, Laura M., J. Leebens-Mack, C. W. DePamphilis, H. Ma, and G. Theissen. 2005. “To B or Not to B a Flower: The Role of DEFICIENS and GLOBOSA Orthologs in the Evolution of the Angiosperms.” Journal of Heredity 96 (3): 225–40. doi:10.1093/jhered/esi033.

Zhang, Gengyun, Xin Liu, Zhiwu Quan, Shifeng Cheng, Xun Xu, Shengkai Pan, Min Xie, et al. 2012. “Genome Sequence of Foxtail Millet (Setaria Italica) Provides Insights into Grass Evolution and Biofuel Potential.” Nature Biotechnology 30 (6): 549–54. doi:10.1038/nbt.2195.

Zhang, Zhen, Haihong Shang, Yuzhen Shi, Long Huang, Junwen Li, Qun Ge, Juwu Gong, et al. 2016. “Construction of a High-Density Genetic Map by Specific Locus Amplified Fragment Sequencing (SLAF-Seq) and Its Application to Quantitative Trait Loci (QTL) Analysis for Boll Weight in Upland Cotton (Gossypium Hirsutum).” BMC Plant Biology 16 (79). BMC Plant Biology: 1–18. doi:10.1186/s12870-016-0741-4.

Zhao, Xinxin, Linkai Huang, Xinquan Zhang, Jianping Wang, Defei Yan, Ji Li, Lu Tang, Xiaolong Li, and Tongwei Shi. 2016. “Construction of High-Density Genetic Linkage Map and Identification of Flowering-Time QTLs in Orchardgrass Using SSRs and SLAF-Seq.” Scientific Reports 6 (29345): 1–10. doi:10.1038/srep29345.

Zhou, Liyuan, Shouye Liu, Weixun Wu, Daibo Chen, Xiaodeng Zhan, Aike Zhu, Yingxin Zhang, et al. 2016. “Dissection of Genetic Architecture of Rice Plant Height and Heading Date by Multiple-Strategy-Based Association Studies.” Scientific Reports 6 (29718). Nature Publishing Group: 1–10. doi:10.1038/srep29718.

206

Zhou, Sifan, Chengran Wang, Taylor F. Frazier, Haidong Yan, Peilin Chen, Zhihong Chen, Linkai Huang, et al. 2018. “The First Illumina-Based de Novo Transcriptome Analysis and Molecular Marker Development in Napier Grass (Pennisetum Purpureum).” Molecular Breeding 38 (95). Molecular Breeding: 1– 14. doi:10.1007/s11032-018-0852-8.

Zhou, Wencai, Zaixiang Tang, Jing Hou, Nan Hu, and Tongming Yin. 2015. “Genetic Map Construction and Detection of Genetic Loci Underlying Segregation Distortion in an Intraspecific Cross of Populus Deltoides.” PLoS ONE, 1–13. doi:10.1371/journal.pone.0126077.

Zhou, Xiaojing, Youlin Xia, Xiaoping Ren, Yulin Chen, Li Huang, Shunmou Huang, Boshou Liao, Yong Lei, Liyin Yan, and Huifang Jiang. 2014. “Construction of a SNP-Based Genetic Linkage Map in Cultivated Peanut Based on Large Scale Marker Development Using next-Generation Double-Digest Restriction-Site- Associated DNA Sequencing (ddRADseq).” BMC Genomics 15 (351): 1–14. doi:10.1186/1471-2164-15-351.

Zub, H. W., S. Arnoult, and M. Brancourt-Hulmel. 2011. “Key Traits for Biomass Production Identified in Different Miscanthus Species at Two Harvest Dates.” Biomass and Bioenergy 35. Elsevier Ltd: 637–51. doi:10.1016/j.biombioe.2010.10.020.

207

BIOGRAPHICAL SKETCH

Dev Paudel received his Ph.D. in Agronomy from the University of Florida (UF) in the fall of 2018. His research focuses on utilizing traditional plant breeding approaches and modern bioinformatic and molecular techniques in the genetic improvement of napiergrass (elephant grass) and its interspecific hybrids with pearl millet. This will eventually help in sustainability of forage and biofuel feedstock.

Dev received his MS degree in Crop Science from Texas Tech University,

Lubbock, Texas, USA where his research focused on evaluating the potential of new testing methods for cotton breeding. Plant breeders, ginners, farmers, and spinning mills can use the information obtained from his research to make informed decisions for increased profitability in the premium yarn market. After his MS degree, he worked at the Texas A&M AgriLife Research, Pecos, Texas where he optimized nutrient media for algae production. He has a BS degree in Agriculture from Tribhuvan University, Nepal.

208