Molecular Phylogenetics and Evolution 85 (2015) 230–237
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev
A mitochondrial genome phylogeny of owlet moths (Lepidoptera: Noctuoidea), and examination of the utility of mitochondrial genomes for lepidopteran phylogenetics ⇑ Xiushuai Yang a,1, Stephen L. Cameron b,1, David C. Lees c,1, Dayong Xue a, Hongxiang Han a, a Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, Beijing 100101, China b Earth, Environmental & Biological Sciences School, Science & Engineering Faculty, Queensland University of Technology, GPO Box 2434, Brisbane, QLD 4001, Australia c Department of Zoology, University of Cambridge, Cambridge CB2 3EJ, United Kingdom article info abstract
Article history: A phylogenetic hypothesis for the lepidopteran superfamily Noctuoidea was inferred based on the com- Received 15 April 2014 plete mitochondrial (mt) genomes of 12 species (six newly sequenced). The monophyly of each noctuoid Revised 27 January 2015 family in the latest classification was well supported. Novel and robust relationships were recovered at Accepted 6 February 2015 the family level, in contrast to previous analyses using nuclear genes. Erebidae was recovered as sister to Available online 17 February 2015 (Nolidae + (Euteliidae + Noctuidae)), while Notodontidae was sister to all these taxa (the putatively basalmost lineage Oenosandridae was not included). In order to improve phylogenetic resolution using Keywords: mt genomes, various analytical approaches were tested: Bayesian inference (BI) vs. maximum likelihood Noctuoidea (ML), excluding vs. including RNA genes (rRNA or tRNA), and Gblocks treatment. The evolutionary signal Mitochondrial genome Phylogeny within mt genomes had low sensitivity to analytical changes. Inference methods had the most significant MAFFT influence. Inclusion of tRNAs positively increased the congruence of topologies, while inclusion of rRNAs Gblocks resulted in a range of phylogenetic relationships varying depending on other analytical factors. The two Gaps Gblocks parameter settings had opposite effects on nodal support between the two inference methods. The relaxed parameter (GBRA) resulted in higher support values in BI analyses, while the strict parameter (GBDH) resulted in higher support values in ML analyses. Ó 2015 Elsevier Inc. All rights reserved.
1. Introduction et al., 2008; Zwick et al., 2011). Surprisingly, the relationships of Bombycidae, Saturniidae, and Sphingidae are still contentious The application of molecular markers in phylogenetics has (e.g. Timmermans et al., 2014), although Breinholt and Kawahara made a tremendous contribution to our knowledge of the evolu- (2013) found phylogenomic evidence for a sister relationship of tion of life on Earth (Delsuc et al., 2005; Wheat and Wahlberg, the last two. The most recent transcriptomic analysis using 46 lepi- 2013). Inference of phylogenetic relationships within Lepidoptera dopteran taxa recovered a novel placement of butterflies and (butterflies and moths), as well their macroevolutionary history established a preliminary framework to understand the evolution (Wahlberg et al., 2013), have been greatly improved by a series of butterflies and moths (Kawahara and Breinholt, 2014). of studies based on an ever-increasing numbers of molecular Noctuoidea is the biggest superfamily of Lepidoptera, number- markers, largely nuclear protein-coding genes (PCGs) (Mutanen ing over 42,400 species (Nieukerken et al., 2011). Many species et al., 2010; Regier et al., 2013). For example, a comprehensive phy- are pests of considerable importance in agriculture and forestry, logenetic study of Bombycoidea (silkworms and hawkmoths) such as armyworms (Spodoptera spp.), bollworms (Helicoverpa based on a combination of up to 25 nuclear genes (19.3 kb total), spp.) and cutworms (Agrotis spp.) (Zhang, 1994). The superfamily the largest to date for a lepidopteran superfamily, found stronger also includes the tiger moths (Arctiinae), one of the more attractive nodal support for many more subclades (families and subfamilies) groups of day-flying moths. The taxonomic history of Noctuoidea is than either a parallel 5-gene analyses or previous studies (Regier complicated, but at present six families are recognized: Oenosan- dridae, Notodontidae, Erebidae, Euteliidae, Nolidae and Noctuidae (Nieukerken et al., 2011; Zahiri et al., 2011). Several other well-
⇑ Corresponding author. known noctuoid groups, including the Lymantriinae and Arctiinae, E-mail address: [email protected] (H. Han). previously accorded family rank are now treated as erebid sub- 1 These authors contributed equally to this work. families (Zahiri et al., 2012). http://dx.doi.org/10.1016/j.ympev.2015.02.005 1055-7903/Ó 2015 Elsevier Inc. All rights reserved. X. Yang et al. / Molecular Phylogenetics and Evolution 85 (2015) 230–237 231
Phylogenetic studies of Noctuoidea have followed the trend of (Barbut and Lalanne-Cassou, 2011; Holloway, 2011). Although increased reliance on molecular markers over morphological char- some potential autapomorphies were proposed in a recent study, acters during the past few decades. The monophyly of Noctuoidea such as the loss of a muscle in the male genitalia, none was indis- (sensu lato) is supported by the presence of a metathoracic tympa- putable and they were absent in some subfamilies or some taxa nal organ, which is regarded as uniquely gained apomorphic char- (Minet et al., 2012). Despite rapid acceptance by a range of lepi- acter (Miller, 1991). Speidel et al. (1996) conducted an analysis of dopteran systematists, the higher-level phylogenetic relationships 30 subfamilies based on 10 morphological characters and con- of Noctuoidea were once again potentially contentious. firmed the monophyly of Noctuidae, but found poor resolution Complete mitochondrial (mt) genomes have been increasingly among subordinate groups. Kitching et al. (1998) extended used to address phylogenetic questions where multi-gene analyses Speidel’s (1996) dataset with larval characters for resolving the have been either unresolved or poorly supported. They have been monophyly of the families Arctiidae, Nolidae, Lymantriidae and widely used in invertebrates in general and insects in particular Pantheidae, as well as their interrelationships. Lafontaine and (Cameron, 2014). Within Lepidoptera, multiple studies have used Fibiger found further support for the respective monophyly of mt genomes to address relationships between superfamilies and the ‘trifid’ noctuoids (Oenosandridae and Notodontidae, with an within major families such as Nymphalidae (Tian et al., 2012; apparently 3-branching forewing cubital vein) and ‘quadrifids’ Yang et al., 2009). These studies have demonstrated significant (with a 4-branching forewing cubital vein) based on adult and and novel findings, including the first strong evidence for the larval characters (Fig. 1B and C) (Fibiger and Lafontaine, 2005; non-monophyly of Macrolepidoptera, primarily due to an earlier Lafontaine and Fibiger, 2006). In their system, all quadrifid moths, diverging position of the butterflies (Papilionoidea) than tradition- including the families Arctiidae, Lymantriidae and Aganaidae, were ally thought (Yang et al., 2009), a position subsequently confirmed redefined as subfamilies of Noctuidae, and therefore Noctuidae by analyses of nuclear PCGs (Mutanen et al., 2010; Regier et al., (sensu lato) became the largest ( 38,000 of 42,000 species) and 2009). Most recently, a comprehensive phylogenetic study of Lepi- most complex noctuoid family (48 subfamilies and 79 tribes). doptera was conducted using mt genomes of 106 species from 24 Molecular studies of Noctuoidea were initially based on one or subfamilies (Timmermans et al., 2014). The topologies are highly two genes, and limited taxon sampling (three families, 29–49 spe- congruent with prior nuclear and/or morphological studies. cies) (Fang et al., 2000; Mitchell et al., 1997, 2000; Weller et al., Various analytical approaches using insect mt genomes to 1994). Mitchell et al. (2000, 2006) initially conducted comprehen- resolve phylogeny have been tested (Cameron, 2014); however, sive systematic analysis of Noctuidae based on two nuclear genes optimal approaches appear to vary between insect lineages. For and a wide range of taxon sampling. Besides finding strong support example, the inclusion of third codon positions positively affected for the monophyly of its subfamilies, the ‘‘LAQ’’ clade was proposed phylogenetic resolution in Orthoptera, Lepidoptera and Diptera for the first time. Lymantriidae and Arctiidae became subordinate (Cameron et al., 2007; Fenn et al., 2008; Yang et al., 2013), while subfamilies within the quadrifid noctuids (Fig. 1D). Mitchell’s the exclusion of third codon positions performed better in Dicty- research, in concert with that of Lafontaine and Fibiger (Fibiger optera and Hymenoptera (Cameron et al., 2012; Dowton et al., and Lafontaine, 2005), has greatly clarified several hitherto elusive 2009). So reporting sensitivity tests for partitioning is of interest relationships within Noctuoidea. However, the support values for different lineages and at different taxonomic scales. In this above the subfamily level were low, possibly resulting from a sam- study, we present six new noctuoid mt genomes, including repre- pling biased towards ‘trifine’ species (with an apparently 3-branch- sentatives of two newly sequenced families (Euteliidae and Noli- ing hindwing cubital vein), with relatively few ‘quadrifines’ (with a dae), and three additional erebid subfamilies. We conducted 4-branching hindwing cubital vein), and/or a result of the limits of phylogenetic analyses of mt genomes from 57 lepidopteran taxa. phylogenetic resolution from the two nuclear genes sequenced. Our aim was to address the infraordinal relationships within Lepi- Recently however, Zahiri et al. (2011) conducted the most exten- doptera, particularly the family-level relationships within Noc- sive molecular phylogenetic analyses of Noctuoidea to date, based tuoidea and subfamily relationships within Erebidae (the largest both on combined sequences of eight PCGs (nuclear and mitochon- noctuoid family) using mt genome data. In addition, we employ a drial) and on comprehensive sampling of 152 species, which cov- range of analytical approaches to explore the phylogenetic utility ered nearly all of the subfamilies recognized within Noctuoidea. of mt genomes. Zahiri et al.’s (2011) results differed significantly from all previous studies, both morphological and molecular ones (Fig. 1E). The tra- 2. Materials and methods ditional group of quadrifid noctuids was broken up, and the families Nolidae and Erebidae (the last family newly defined to 2.1. Samples and sequencing encompass a still vast clade of some 18 subfamilies and about 24,500 species) were reconstituted and Euteliidae was newly rec- Complete mt genomes were sequenced for six noctuoid species: ognized (previously a noctuid subfamily) (Zahiri et al., 2012). In Asota plana lacteata (Erebidae, Aganainae), Agylla virilis (Erebidae, so doing, they proposed a novel six-family system that has current- Arctiinae), Catocala deuteronympha (Erebidae, Erebinae), Eutelia ly been accepted within synoptic classifications of Lepidoptera adulatricoides (Euteliidae, Euteliinae), Gabala argentata (Nolidae, (Nieukerken et al., 2011). While the datasets of Zahiri et al. Chloephorinae), and Risoba prominens (Nolidae, Risobinae). Three (2011, 2012, 2013a,b) provide strong support for the monophyly legs from each specimen were removed and preserved in 100% of the six families and most of the subfamilies, nodal support for ethanol during collection. Legs were stored at 20 °C until DNA Erebidae + Nolidae and most inter-subfamily relationships was extraction, and voucher specimens were deposited at the Museum not significant. This is despite a sizeable molecular dataset of the Institute of Zoology, Chinese Academy of Sciences. Total (6407 bp) including most of the nuclear genes shown to be most genomic DNA was extracted from the leg muscle tissue with the effective at deep phylogenetic levels within Lepidoptera DNeasy Blood and Tissue Kit (QIAGEN, USA). Mt genomes were (Mutanen et al., 2010; Wahlberg et al., 2009a); further resolution amplified and sequenced as described in our previous study of noctuoid relationships thus requires novel data sources. An (Yang et al., 2013). Generally, primer walking was employed for additional outcome of Zahiri et al. (2011, 2012, 2013a,b) was that long fragments; fragments containing the control region were morphological and molecular studies were contradictory, with no cloned using TOPOÒ TA CloningÒ Kit (Invitrogen, USA), and the morphological support for Erebidae. In practice, the Erebidae is resultant plasmid DNA was isolated using the TIANpure Midi Plas- impossible to recognize without comprehensive faunal knowledge mid Kit (Tiangen, China). The sequence data have been deposited in 232 X. Yang et al. / Molecular Phylogenetics and Evolution 85 (2015) 230–237
Fig. 1. Previous hypotheses of phylogenetic relationships within Noctuoidea: (A) Speidel et al. (1996); note: Noctuidae in this study includes the quadrifine groups; (B) Fibiger and Lafontaine (2005); (C) Lafontaine and Fibiger (2006); (D) Mitchell et al. (2006); (E) Zahiri et al. (2011) and (F) Zahiri et al. (2013b). Family and subfamily names are as Zahiri et al. (2011). Terminals were changed from the original publication to track Zahiri et al. (2011) and allow easy comparison between studies.
GenBank under accession numbers KJ173908 (A. plana lacteata), selection was via Akaike Information Criterion implemented in KJ185131 (E. adulatricoides), KJ396197 (R. prominens), KJ410747 MrModeltest 2.3 (Nylander, 2004). For BI analyses, Ahamus yunna- (G. argentata), KJ364659 (A. virilis), and KJ432280 (C. deuteronympha). nensis (Hepialidae) was selected as the outgroup (Lee et al., 2006). Four independent runs were conducted for 1,000,000 generations, 2.2. Annotation and alignment and each was sampled every 1000 generations. Convergence was determined using Tracer ver. 1.5 (Rambaut and Drummond, Sequence assembly and mt genome annotation followed our 2007) and all analyses had converged within 1,000,000 gen- previous study (Yang et al., 2013). Alignments of individual genes erations. Posterior probabilities over 0.95 were interpreted as were performed using MAFFT (Katoh et al., 2002). Q-INS-i, the strongly-supported (Erixon et al., 2003). For ML analyses, two RNA structural alignment method implemented in MAFFT, was Hepialidae species Ahamus yunnanensis and Thitarodes renzhiensis explored for rRNA and tRNA alignments and PCGs were aligned with were selected as outgroups, and 1000 replicate bootstraps were the default settings (Katoh and Toh, 2008). Three different datasets conducted under GTRCAT. Nodes with a bootstrap percentage were generated to test the effect of data inclusion/exclusion on (BP) of at least 70% were considered well-supported in the ML ana- topology and nodal support: (1) 13PCG123, complete 13 PCGs (all lyses (Hillis and Bull, 1993). RNA genes omitted); (2) 13PCG123tRNA, 13PCG123 + tRNA genes (rRNA genes omitted); (3) 13PCG123LRSR, 13PCG123 + rRNA genes 3. Results (tRNA genes omitted). In order to remove unreliably aligned regions within the data- 3.1. Gene order and variation sets, we used Gblocks (Castresana, 2000) to identify the conserved regions with the two following parameter settings: GBDH, default The six noctuoid species were all completely sequenced, and all parameter settings except allowing gaps at a given position in up of them possessed the typical metazoan mt genome composition of to half of the aligned species; GBRA, relaxed parameter settings 13 PCGs, 22 transfer RNAs, two ribosomal RNAs and a single A + T- (Minimum Number Of Sequence For A Conserved Position: mini- rich region (putative control region) (Fig. 2)(Boore, 1999). The mum; Minimum Number Of Sequence For A Flank Position: mini- A + T rich regions of noctuoid species were routinely short with a mum; Maximum Number Of Contiguous Nonconserved Positions: range of 311–541 bp, as well as a high AT content of 90.6–96.1%. default; Minimum Length Of A Block: minimum; Allowed Gap The mt gene order was identical to other noctuoid species previ- Positions: all). Additionally, datasets analyzed without Gblocks ously published, which possessed the gene rearrangement of treatments were adjusted manually to remove the non- tRNAMet found in many lepidopterans. This character, once sug- homologous regions at the beginning and the end of each aligned gested to be synapomorphic for Lepidoptera (Gong et al., 2012), gene. has recently been found to be restricted to the more derived clade Ditrysia (Cao et al., 2012) and their close relatives Tischerioidea 2.3. Phylogenetic analyses and Palaephatoidea (Timmermans et al., 2014).
Each dataset was analyzed using two inference methods, Baye- 3.2. Problematic mt genomes within the first dataset sian inference (BI) and maximum likelihood (ML). Analyses were performed using MrBayes v 3.2.2 (Ronquist et al., 2012) and A range of preliminary analyses were conducted using the six RAxML-HPC2 v 7.6.3 (Stamatakis, 2006) on XSEDE via the Cyberin- newly sequenced mt genomes in this study plus all complete mt frastructure for Phylogenetic Research (CIPRES) Science gateway genomes of Lepidoptera in GenBank (61 species by March 2013). portal (Miller et al., 2010). Two rRNA, 22 tRNA genes and 13 PCGs However, some taxa were obviously misplaced in all analyses. were treated as 4 partitions, lrRNA, srRNA, a combined partition of For example, in no case was a monophyletic Hesperiidae (Ochlodes the 22 tRNA genes, and a combined partition of 13 PCGs. Model venata and Ctenoptilum vasava) recovered. Instead, all datasets X. Yang et al. / Molecular Phylogenetics and Evolution 85 (2015) 230–237 233
Fig. 2. Mt genome organization and mt genome sizes for the six newly sequenced noctuoid species.
supported the grouping of O. venata and three Lycaenidae species also varied between inference methods. For the M13PCG123ma (Fig. S1). Sequence blasting of the 50 COI of the O. venata mt and 13PCG123tRNAGBRA datasets, Pieridae and Lycaenidae were genome published on GenBank does not closely match other sam- recovered as a sister group (the B2-type relationship is presented ples identified to this species or genus on BOLD (http://www. in Fig. 3) in ML analyses, whereas BI analyses recovered Nym- boldsystems.org)(Ratnasingham and Hebert, 2007). The phalidae and Lycaenidae (the B1-type relationship is presented in monophyly of Hesperiidae is not in dispute (Heikkilä et al., 2012; Fig. 3) as sister groups (these relationships, however, also vary Warren et al., 2008, 2009). Similarly, one of the lycaenid species between datasets for the same inference methods, see below, and (Celastrina hersilia) grouped with Papilionidae. BLAST identification note due to our cutoff for public sequences we did not include a of this sequence suggested it was from another genus (Cyaniris). recently published sequence for Lycaenidae’s widely considered The controversial placements of O. venata and C. hersilia need sister group, Riodinidae). Relationships within families showed further investigation (see below) to determine whether they little sensitivity to inference method, however the relationships are caused by contamination during sequencing, or specimen within Olethreutinae (Tortricidae) and the placement of Libythea misidentification. celtis within Nymphalidae were divergent in the two inference Finally, we found that the divergence between two mt genomes methods, while the placement of Noctuidae within Noctuoidea of Parnassius bremeri (GenBank accession nos. HM243588 and was another example in BI analyses (Fig. 3). NC_014053) reached 4.0%, which was quite unexpected for con- For the different data treatments, we compared the effect of specific taxa. Further alignment analysis showed that the region inclusion of rRNA or tRNA genes in combined analyses of from the tRNA cluster ARNSEF to lrRNA of NC_014053 was the PCGs + rRNA and PCGs + tRNA against PCGs alone. Inclusion of tRNA most divergent region (8.2%), while the rest of the mt genome genes had no effect on tree topology and slight effect on nodal was only 0.5% divergent. Further blasting analysis showed that support when the datasets including third codon positions were the divergent region of NC_014053 had 94% identity to C. hersilia, analyzed using BI (13PCG123tRNAmaGBDH vs. 13PCG123maGBDH while only as much as 87% to other species of Papilionidae. This and 13PCG123tRNAmaGBRA vs. 13PCG123maGBRA). However, suggests that the sequence deposited on GenBank might be a different relationships were recovered within Noctuoidea, Nym- chimera, due to a contig formed of long PCR resulting from more phalidae and Tortricidae when rRNA genes were included than one species. (13PCG123LRSRmaGBDH and 13PCG123LRSRmaGBRA). In BI ana- In the preliminary analyses, most of the problems were within lyses, Euteliidae and Nolidae were recovered as sister groups (the the unpublished data deposited in GenBank, so we conducted D2-type relationship is presented in Fig. 3) in 13PCG123LRSR- analyses (57 taxa) excluding the unpublished data in this study maGBDH and 13PCG123LRSRmaGBRA, while Noctuidae and Nolidae (Table S1). We recommend careful checking and revision before were recovered as a sister group (the D1-type relationship is pre- conducting the phylogenetic analyses with lepidopteran mt sented in Fig. 3) from the other datasets (PCGs + tRNA and PCGs genomes from GenBank, especially those associated with unpub- alone). Diverse interfamily and intrafamily relationships were lished studies. recovered across the different data treatments using ML analyses. The B2-type relationship (Lycaenidae + (Pieridae + Nymphalidae)) 3.3. Methodological effects of various approaches was recovered from PCGs + tRNA (13PCG123tRNAmaGBDH and 13PCG123tRNAmaGBRA), while the B1-type relationship Fourteen independent phylogenetic analyses were performed to ((Lycaenidae +Pieridae) + Nymphalidae) was recovered from test sensitivity to inference methods, data treatment approaches, PCGs + rRNA (13PCG123LRSRmaGBDH and 13PCG123LRSR- and variable region exclusion with Gblocks. A total of eight maGBRA). More diverse relationships were recovered within Nym- different tree topologies were recovered (Figs. 3, S2 and S3). A brief phalidae in ML analyses. 13PCG123tRNAmaGBDH recovered summary of the datasets is displayed in Table 1. L. celtis to be the sister group of the remaining nymphalids (exclud- The two inference methods (BI and ML) resulted in different ing Euploea mulciber) (the C1-type relationship is presented in tree topologies for the same dataset. The monophyly of Yponomeu- Fig. 3), while 13PCG123maGBDH and 13PCG123LRSRmaGBDH toidea was supported by all the BI analyses (although not sig- recovered L. celtis as the sister group of a clade consisting of Calinaga nificantly in 13PCG123maGBDH and 13PCG123LRSRmaGBDH) davidis plus Hipparchia autonoe (the C3-type relationship is whereas as it was never significantly supported in the ML analyses presented in Fig. 3). and the superfamily did not emerge as monophyletic (Plutella Using Gblocks with two different parameter settings appeared xylostella was sister to Apoditrysia) in 13PCG123tRNAmaGBDH to have limited effect on phylogenetic reconstruction under the and 13PCG123LRSRmaGBDH. Relationships within Papilionoidea Bayesian framework. The only difference found was between PCGs 234 X. Yang et al. / Molecular Phylogenetics and Evolution 85 (2015) 230–237
Fig. 3. Summary of different phylogenetic relationships recovered in this study. The entire tree was inferred from the Bayesian analysis of the 13PCG123maGBRA dataset, and different types of relationships recovered in the Bayesian and maximum likelihood analyses from the other datasets are displayed on the left.
Table 1 vs. 13PCG123tRNAmaGBRA and 13PCG123LRSRmaGBDH vs. Brief summary of the datasets established from the mt genomes of 57 species. 13PCG123LRSRmaGBRA). In addition, Gblocks treatments moder- Dataset name Dataset size (bp) ately affected topology when tRNA or rRNA genes were included. M13PCG123ma 11,263 The strict parameter Gblocks (GBDH) ML analyses failed to recover 13PCG123maGBDH 10,959 the monophyly of Yponomeutoidea from PCGs + tRNA and 13PCG123maGBRA 11,258 PCGs + rRNA datasets (Fig. S3). These results indicate that ML 13PCG123tRNAmaGBDH 12,159 analysis is more sensitive to gaps than BI analysis. 13PCG123tRNAmaGBRA 12,947 13PCG123LRSRmaGBDH 12,558 13PCG123LRSRmaGBRA 13,604 3.4. Phylogeny Abbreviation and explanation: ‘M’, manually (manually adjusted the genes after alignment); ‘ma’, MAFFT (genes were aligned by MAFFT); ‘GBDH’, Gblocks masking Phylogenetic analyses were performed using 57 complete mt with the strict parameter settings; ‘GBRA’, Gblocks masking with the relaxed genomes, representing eight lepidopteran superfamilies. The parameter settings; ‘LR’, large subunit ribosomal RNA; ‘SR’, small subunit ribosomal monophyly of each superfamily was generally well supported, RNA. except for two ML analyses, in which Yponomeutoidea became paraphyletic, with Leucoptera malifoliella sister to Plutella xylostel- la + other six ingroup superfamilies. Relationships inferred at the and PCGs + tRNA. Datasets with the strict parameter settings superfamily level were highly consistent among all analyses. How- (13PCG123maGBDH and 13PCG123tRNAmaGBDH) recovered the ever, we do not expect a robust phylogenetic relationship at the C2-type relationship within Nymphalidae (C. davidis plus H. auto- superfamily level due to only a small proportion of the lepidopter- noe sister to all Nymphalidae except E. mulciber), while datasets an superfamilies being represented in this study. The monophyly with the relaxed parameter settings (13PCG123maGBRA and of Yponomeutoidea is well supported after the removal of certain 13PCG123tRNAmaGBRA) recovered the C1-type relationship (see problematic taxa. Relationships within Bombycoidea, Geometroi- above). For the three PCGs datasets, the dataset generated by dea and Pyraloidea were almost identical to our previous study Gblocks with the relaxed parameter settings (13PCG123maGBRA) (Yang et al., 2013). Relationships within Olethreutinae (Tortricidae, did not influence topology but generally increased nodal support Tortricoidea) were sensitive to different analytical approaches values when compared with the dataset manually adjusted (Fig. 3, A1–A4). However, we found some novel relationships with- (M13PCG123ma). In ML analyses, Gblocks also had limited influ- in Noctuoidea and Papilionoidea. ence among the three PCGs datasets. However, the PCGs dataset Novel relationships were found for the Noctuoidea families with the strict parameter settings of Gblocks (13PCG123maGBDH) (Fig. 3) compared to previous molecular phylogenies of this super- inferred a tree with less weakly supported nodes than the one with family. The redefined family Noctuidae grouped with the recently the relaxed parameter settings (13PCG123maGBRA), and the same recognized family Euteliidae. Noctuidae + Euteliidae was sister to pattern was found for the other datasets (13PCG123tRNAmaGBDH the reconstituted family Nolidae, and the recently reconstituted X. Yang et al. / Molecular Phylogenetics and Evolution 85 (2015) 230–237 235 family Erebidae was sister to the clade of (Nolidae + (Euteli- some datasets, indicating that this incongruence is the result of idae + Noctuidae)). Notodontidae was the sister group of the other more comprehensive taxon sampling. Therefore, it will be impor- four noctuoid families. Within Erebidae, only four subfamilies were tant to employ different inference methods to evaluate the phylo- analyzed, and their relationships were as follows: (Arctiinae + (Ere- genetic signal as ever larger taxon datasets are analyzed. binae + (Lymantriinae + Aganainae))). Phylogenetic relationships Whether to use gene or codon exclusion is an ongoing debate in within Noctuoidea were consistent among most of the datasets relation to the accuracy of phylogenetic analyses using mt genomes. except those including rRNA genes (13PCG123LRSRmaGBDH and Our previous study indicated that it was not essential to exclude 13PCG123LRSRmaGBRA) in BI analyses (Euteliidae was sister to third codon positions in reconstructing lepidopteran phylogeny Nolidae, the D2-type relationship Fig. 3), and received strong sup- (Yang et al., 2013). In this study, we further tested the effect of gene port from all PCG123 datasets in the BI analyses and in most of ML exclusion by comparing the combined analyses of PCGs + rRNA/ analyses (Figs. 3 and S2). tRNA with analyses of PCGs alone. rRNAs and tRNAs are often In the present analyses, a total of 21 Papilionoidea species were excluded in phylogenetic analyses of insects. However, analyses of included representing five families and 12 subfamilies. However, dipteran mt genomes suggested including rRNAs and tRNAs two controversial nodes within Papilionoidea that have been found improved resolution and nodal support (Cameron et al., 2007). Simi- in previous analyses of mt genomes still remained (Yang et al., 2013). lar conclusions were also obtained by other studies of Orthoptera In particular, the position of L. celtis was unstable with respect to and Neuropterida (Cameron et al., 2009; Fenn et al., 2008). In our Danainae (represented by E. mulciber) and Calinaginae + Satyrinae analyses, the inclusion of tRNAs positively increased the congruence (represented by C. davidis and H. autonoe respectively), and three dif- of topologies, indicating tRNAs contribute signal to phylogenetic ferent possible relationships were inferred (Fig. 3). The C1-type rela- analyses. By contrast, inclusion of rRNAs resulted in more diverse tionship (Danainae + (Libytheinae + remaining Nymphalidae)) and poorly supported phylogenetic relationships compared with received the highest support from both BI and ML analyses. The other datasets. Three candidate explanations might be responsible other unstable relationships were among Pieridae, Lycaenidae and for noisy signals in rRNAs. First, rRNAs have long been a challenge Nymphalidae. The B1-type, more morphologically intuitive to align accurately. Although incorporating structural information relationship (Pieridae + (Lycaenidae + Nymphalidae)) was well sup- into alignment promises to improve the accuracy, this is difficult ported in most of the BI analyses (Fig. S2). However, in ML analyses, to apply for more distantly related taxa (Cameron et al., 2009). the B1-type relationship received strong support just in Another possible reason is that rRNA genes have much higher levels 13PCG123LRSRmaGBRA, and the B2-type relationship ((Pieridae + of homoplasy than either PCGs or tRNA genes (Cameron et al., 2007). Lycaenidae) + Nymphalidae)) was well supported in three datasets A further reason is that compensatory substitutions in paired (M13PCG123ma, 13PCG123maGBDH, and 13PCG123tRNA- regions of RNA genes are not directly accounted for by the phyloge- maGBDH) (Fig. S3). netic inference methods unless explicitly modeled, such as the dou- blet model (Schöniger and Von Haeseler, 1994) used by Letsch et al. 4. Discussion (2010). Therefore, substitutions at each site are assumed to be inde- pendent of one another possibly resulting in conflicting signal to the 4.1. Methodological effects of various approaches PCGs partition (Hudelot et al., 2003). Gblocks was used to improve signal-to-noise ratio by removing In our previous study (Yang et al., 2013), we have tested the highly variable regions, caused by poor alignment or inaccurately effect of inference and alignment methods on phylogenetic recon- defined gene boundaries. For example, most CO1 genes in Lepi- struction, including or excluding the 3rd codon positions from doptera have been predicted to start with the nonstandard start PCGs. Those results showed that inference using BI and ML, align- codon-CGA, while others were predicted to start with the standard ment using Muscle and MAFFT, and including third codon positions start codon-ATN (Fig. S4). Therefore, the relatively arbitrary defini- all performed better than maximum parsimony analysis, Clustal tion of a gene boundary might or might not include variable alignment and excluding third codon positions, respectively. The regions at the gene’s 50 end. Similar problems exist in many other most recent research on Noctuidae based on COI and seven nuclear genes, such as ND2, ND4, ND4L, and ND5. However, recent com- genes arrived at a similar conclusion that removing the third codon parative analyses suggest that removing variable regions has a positions resulted in unstable phylogeny and weakened support negative impact on phylogenetic accuracy (Dessimoz and Gil, values (Zahiri et al., 2013b). In the present study, we extended 2010). We employed two different parameter settings of Gblocks our exploration of mt genome phylogenetic utility by testing two in our analyses. The results showed that the relaxed parameters, inference methods (BI and ML), the inclusion of rRNA or tRNA GBRA, resulted in variant relationships and poor support values genes and Gblocks removal of variable regions. Overall, lepidopter- in ML analyses, while the strict parameter, GBDH, had a beneficial an mt genomes have low sensitivity to these various analytical effect, increasing nodal support. By contrast, the two parameter approaches, and thus are of high utility in resolving higher-level settings both resulted in more variable topologies and poorer nodal relationships between and within the lepidopteran superfamilies support in BI analyses. A similar conclusion was drawn by Dwivedi (Timmermans et al., 2014; Yang et al., 2013). We found that infer- and Gadagkar (2009), and they suggested that the possible reason ence method was the most significant factor, which is consistent for the opposite effects was due to the different criteria for treating with a previous study of neuropteridan mt genomes (Cameron gaps as missing data in the different inference methods. et al., 2009). In all instances, BI analyses recovered a more consis- tent tree topology across different datasets, and nodal support was 4.2. Phylogeny impressively high for the majority of nodes. However, it has been suggested in several recent studies that posterior probabilities The phylogeny and classification of Noctuoidea has experienced are generally too liberal (Cummings et al., 2003; Simmons et al., a chaotic history, changing rapidly with the implementation of 2004). Thus, caution should be used if BI analysis is solely applied. molecular markers and the discovery of morphological synapo- In contrast, ML analyses revealed many more topological differ- morphies. The status of Noctuidae (sensu stricto), including all tri- ences among varying datasets. Additionally, for none of the seven fine noctuids and some ‘pseudoquadrifine’ noctuids, tends to be datasets was the same topology recovered between the two infer- consistent with recent morphological and molecular researches ence methods. This is different from our previous study where the (Fibiger and Lafontaine, 2005; Lafontaine and Fibiger, 2006; two inference methods recovered identical tree topologies for Mitchell et al., 2006; Zahiri et al., 2011, 2013b). Noctuoidea with 236 X. Yang et al. / Molecular Phylogenetics and Evolution 85 (2015) 230–237 a quadrifid forewing and a quadrifine hindwing, consisting of the 31372176) and the Key Laboratory of the Zoological Systematics most species-rich groups such as Lymantriinae and Arctiinae, are and Evolution of the Chinese Academy of Sciences (No. also the lineages most often subject to change between different O529YX5105). David Lees was funded by ERC Grant #250325 (Bel- classification schemes. In addition, various noctuoid classification gium) during the writing of this study. Stephen Cameron was fund- schemes have not tended towards a consensus of relationships at ed by an Australian Research Council Future Fellowship family or subfamily levels. To date, the most comprehensive mole- (FT120100746). cular phylogenetic analyses of Noctuoidea were conducted by Zahiri et al. (2011), from which a novel classification system for this giant superfamily was proposed. Support values for each of Appendix A. Supplementary material the six major clades were strong, although least so (0.96 posterior probability) for Erebidae, and monophyly of each erebid subfamily Supplementary data associated with this article can be found, in was moderately to strongly supported (Zahiri et al., 2011). Zahiri the online version, at http://dx.doi.org/10.1016/j.ympev.2015.02. et al.’s (2011) family system was however well supported by our 005. mitochondrial phylogeny, despite our limited taxon sampling. Novel and robust relationships were, however, recovered at both References family and subfamily levels that are in conflict with Zahiri et al. (2011). In our analyses, the position of Notodontidae as sister to Barbut, J., Lalanne-Cassou, B., 2011. Description of a new species of Eulepidotis the remaining noctuoid families and the sister grouping of Noctu- Hübner, 1823, discovered in the natural reserve of the Trinité (French Guiana) idae plus Euteliidae were congruent with the ML analyses in Zahiri (Lepidoptera, Erebidae, Eulepidotinae). Bull. Soc. Entomol. Fr. 116, 181–184. Boore, J.L., 1999. Animal mitochondrial genomes. Nucleic Acids Res. 27, 1767–1780. et al. (2011). In contrast, Nolidae was found to have a closer rela- Breinholt, J.W., Kawahara, A.Y., 2013. Phylotranscriptomics: saturated third codon tionship with the clade (Noctuidae + Euteliidae), rather than with positions radically influence the estimation of trees based on next-gen data. Erebidae as suggested by the BI analyses of Zahiri et al. (2013b). Genome Biol. Evol. 5, 2082–2092. Cameron, S.L., 2014. Insect mitochondrial genomics: implications for evolution and A different relationship between Noctuidae and quadrifine noctu- phylogeny. Annu. Rev. Entomol. 59, 95–117. ids was recovered in our BI analyses of datasets 13PCG123tRNA- Cameron, S.L., Lambkin, C.L., Barker, S.C., Whiting, M.F., 2007. A mitochondrial maGBRA and 13PCG123LRSRmaGBRA, with Noctuidae recovered genome phylogeny of Diptera: whole genome sequence data accurately resolve as the sister group to the clade (Euteliidae + Nolidae). In the pre- relationships over broad timescales with high precision. Syst. Entomol. 32, 40– 59. sent study, Arctiinae was recovered as the sister of the remaining Cameron, S.L., Sullivan, J., Song, H., Miller, K.B., Whiting, M.F., 2009. A mitochondrial erebid subfamilies, which is consistent with the recent analyses genome phylogeny of the Neuropterida (lace-wings, alderflies and snakeflies) based on 19 nuclear PCGs in Regier et al. (2013). Interestingly, Aga- and their relationship to the other holometabolous insect orders. Zool. Scr. 38, 575–590. nainae and Lymantriinae were here grouped together for the first Cameron, S.L., Lo, N., Bourguignon, T., Svenson, G.J., Evans, T.A., 2012. A time, whereas Aganainae was found to have a closer relationship mitochondrial genome phylogeny of termites (Blattodea: Termitoidae): robust with Arctiinae in recent molecular and morphological studies support for interfamilial relationships and molecular synapomorphies define major clades. Mol. Phylogen. Evol. 65, 163–173. (Figs. 1 and 3)(Fibiger and Lafontaine, 2005; Zahiri et al., 2012). Cao, Y.Q., Ma, C., Chen, J.Y., Yang, D.R., 2012. The complete mitochondrial genomes Zahiri et al. (2011) proposed that trifine noctuids evolved from quad- of two ghost moths, Thitarodes renzhiensis and Thitarodes yunnanensis: the rifines multiple times, due to hindwing vein reduction in some quad- ancestral gene arrangement in Lepidoptera. BMC Genom. 13, 276. Castresana, J., 2000. Selection of conserved blocks from multiple alignments for rifines and the position of quadrifine species within some trifine their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552. subfamilies. This proposal is also supported by our mt genome phy- Cummings, M.P., Handley, S.A., Myers, D.S., Reed, D.L., Rokas, A., Winka, K., 2003. logenies as the trifine groups Arctiinae and Nolidae never formed a Comparing bootstrap and posterior probability values in the four-taxon case. Syst. Biol. 52, 477–487. monophyletic group; however, the multiple origins of these wing Delsuc, F., Brinkmann, H., Philippe, H., 2005. Phylogenomics and the reconstruction venation patterns in Noctuidae still needs more research. These of the tree of life. Nat. Rev. Genet. 6, 361–375. questions can probably be resolved by extending both taxon sam- Dessimoz, C., Gil, M., 2010. Research phylogenetic assessment of alignments reveals pling (especially additional representatives of subfamilies/tribes neglected tree signal in gaps. Genome Biol. 11, R37. Dowton, M., Cameron, S.L., Austin, A.D., Whiting, M.F., 2009. Phylogenetic within the massive groups Erebidae and Noctuidae families) and approaches for the analysis of mitochondrial genome sequence datain the the number of molecular markers. The combination of mt genomes Hymenoptera – a lineage with both rapidly and slowly evolving mitochondrial with nuclear genes (at least when a limited proportion of the nuclear genomes. Mol. Phylogenet. Evol. 52, 512–519. Dwivedi, B., Gadagkar, S.R., 2009. Phylogenetic inference under varying proportions genome or transcriptome is used) could be an effective way to refine of indel-induced alignment gaps. BMC Evol. Biol. 9, 211. the classification of noctuoid families and test the robustness of Erixon, P., Svennblad, B., Britton, T., Oxelman, B., 2003. Reliability of Bayesian interfamily relationships. A stable and informative classification of posterior probabilities and bootstrap frequencies in phylogenetics. Syst. Biol. 52, 665–673. Noctuoidea will ultimately help interpret and/or discover relevant Fang, Q.Q., Mitchell, A., Regier, J.C., Mitter, C., Friedlander, T.P., Poole, R.W., 2000. morphological autapomorphies, especially for Erebidae and Nolidae. Phylogenetic utility of the nuclear gene dopa decarboxylase in noctuoid moths Many previous studies have demonstrated great differences in (Insecta: Lepidoptera: Noctuoidea). Mol. Phylogen. Evol. 15, 473–486. Fenn, J.D., Song, H., Cameron, S.L., Whiting, M.F., 2008. A preliminary mitochondrial phylogenetic signal between mitochondrial and nuclear genes genome phylogeny of Orthoptera (Insecta) and approaches to maximizing (Lin and Danforth, 2004; Rubinoff and Holland, 2005; Shaw, phylogenetic signal found within mitochondrial genome data. Mol. Phylogenet. 2002; Wahlberg et al., 2009b). However, the mitochondrial phy- Evol. 49, 59–68. Fibiger, M., Lafontaine, J.D., 2005. A review of the higher classification of the logeny of Noctuoidea in the present study shows promising con- Noctuoidea (Lepidoptera) with special reference to the Holarctic fauna. gruence with the previous, almost entirely nuclear PCG-based Esperiana 11, 7–92. studies (Zahiri et al., 2011), suggesting that greater integration of Gong, Y.J., Shi, B.C., Kang, Z.J., Zhang, F., Wei, S.J., 2012. The complete mitochondrial nuclear and mt genomes will indeed lead to an improvement in genome of the oriental fruit moth Grapholita molesta (Busck) (Lepidoptera: Tortricidae). Mol. Biol. Rep. 39, 2893–2900. our knowledge of insect evolution and phylogeny, although the Heikkilä, M., Kaila, L., Mutanen, M., Peña, C., Wahlberg, N., 2012. Cretaceous origin importance of growing amounts of nuclear data (e.g. Kawahara and repeated tertiary diversification of the redefined butterflies. Proc. R. Soc. and Breinholt, 2014) cannot be understated. Biol. Sci. Ser. B 279, 1093–1099. Hillis, D.M., Bull, J.J., 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst. Biol. 42, 182–192. Acknowledgments Holloway, J., 2011. The moths of Borneo: families Phaudidae, Himantopteridae and Zygaenidae; revised and annotated checklist. Malay. Nat. J., 63 Hudelot, C., Gowri-Shankar, V., Jow, H., Rattray, M., Higgs, P.G., 2003. RNA-based This study was supported by the grants from the National Nat- phylogenetic methods: application to mammalian mitochondrial RNA ural Science Foundation of China (Nos. 31272288, 31071952, and sequences. Mol. Phylogen. Evol. 28, 241–252. X. Yang et al. / Molecular Phylogenetics and Evolution 85 (2015) 230–237 237
Katoh, K., Toh, H., 2008. Recent developments in the MAFFT multiple sequence phylogenetic inference and model choice across a large model space. Syst. Biol. alignment program. Brief. Bioinform. 9, 286–298. 61, 539–542. Katoh, K., Misawa, K., Kuma, K.i., Miyata, T., 2002. MAFFT: a novel method for rapid Rubinoff, D., Holland, B.S., 2005. Between two extremes: mitochondrial DNA is multiple sequence alignment based on fast Fourier transform. Nucleic Acids neither the panacea nor the nemesis of phylogenetic and taxonomic inference. Res. 30, 3059–3066. Syst. Biol. 54, 952–961. Kawahara, A.Y., Breinholt, J.W., 2014. Phylogenomics provides strong evidence for Schöniger, M., Von Haeseler, A., 1994. A stochastic model for the evolution of relationships of butterflies and moths. Proc. R. Soc. Biol. Sci. Ser. B 281, autocorrelated DNA sequences. Mol. Phylogen. Evol. 3, 240–247. 20140970. Shaw, K.L., 2002. Conflict between nuclear and mitochondrial DNA phylogenies of a Kitching, I.J., Rawlins, J.E., Kristensen, N.P., 1998. The Noctuoidae. Handbook of recent species radiation: what mtDNA reveals and conceals about modes of Zoology, vol. IV Arthropoda: Insecta. Lepidoptera, Moths and Butterflies, vol. 1: speciation in Hawaiian crickets. Proc. Natl. Acad. Sci. 99, 16122–16127. Evolution, Systematics and Biogeography, pp. 355–402. Simmons, M.P., Pickett, K.M., Miya, M., 2004. How meaningful are Bayesian support Lafontaine, J.D., Fibiger, M., 2006. Revised higher classification of the Noctuoidea values? Mol. Biol. Evol. 21, 188–199. (Lepidoptera). Can. Entomol. 138, 610–635. Speidel, W., Fänger, H., Naumann, C., 1996. The phylogeny of the Noctuidae Lee, E.S., Shin, K.S., Kim, M.S., Park, H., Cho, S.W., Kim, C.B., 2006. The mitochondrial (Lepidoptera). Syst. Entomol. 21, 219–252. genome of the smaller tea tortrix Adoxophyes honmai (Lepidoptera: Tortricidae). Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic Gene 373, 52–57. analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688– Letsch, H.O., Kück, P., Stocsits, R.R., Misof, B., 2010. The impact of rRNA secondary 2690. structure consideration in alignment and tree reconstruction: simulated data Tian, L.L., Sun, X.Y., Chen, M., Gai, Y.H., Hao, J.S., Yang, Q., 2012. Complete and a case study on the phylogeny of hexapods. Mol. Biol. Evol. 27, 2507–2521. mitochondrial genome of the five-dot sergeant Parathyma sulpitia Lin, C.P., Danforth, B.N., 2004. How do insect nuclear and mitochondrial gene (Nymphalidae: Limenitidinae) and its phylogenetic implications. Zool. Res. 33, substitution patterns differ? insights from Bayesian analyses of combined 133–143. datasets. Mol. Phylogen. Evol. 30, 686–702. Timmermans, M.J., Lees, D.C., Simonsen, T.J., 2014. Towards a mitogenomic Miller, J.S., 1991. Cladistics and classification of the Notodontidae (Lepidoptera, phylogeny of Lepidoptera. Mol. Phylogen. Evol. 79, 169–178. Noctuoidea) based on larval and adult morphology. Bull. Am. Mus. Nat. Hist., 1– Wahlberg, N., Leneveu, J., Kodandaramaiah, U., Peña, C., Nylin, S., Freitas, A.V., 221 Brower, A.V., 2009a. Nymphalid butterflies diversify following near demise at Miller, M.A., Pfeiffer, W., Schwartz, T., 2010. Creating the CIPRES Science Gateway the Cretaceous/Tertiary boundary. Proc. R. Soc. Biol. Sci. Ser. B 276, 4295–4302. for inference of large phylogenetic trees. In: Gateway Computing Environments Wahlberg, N., Weingartner, E., Warren, A., Nylin, S., 2009b. Timing major conflict Workshop (GCE). IEEE, pp. 1–8. between mitochondrial and nuclear genes in species relationships of Polygonia Minet, J., Barbut, J., Lalanne-Cassou, B., 2012. Les Noctuelles: classification et clé de butterflies (Nymphalidae: Nymphalini). BMC Evol. Biol. 9, 92. détermination des familles (Lepidoptera: Noctuoidea). Alexanor 25, 131–151. Wahlberg, N., Wheat, C.W., Peña, C., 2013. Timing and patterns in the taxonomic Mitchell, A., Cho, S., Regier, J.C., Mitter, C., Poole, R.W., Matthews, M., 1997. diversification of Lepidoptera (butterflies and moths). PLoS ONE 8, e80875. Phylogenetic utility of elongation factor-1 alpha in Noctuoidea (Insecta: Warren, A.D., Ogawa, J.R., Brower, A.V., 2008. Phylogenetic relationships of Lepidoptera): the limits of synonymous substitution. Mol. Biol. Evol. 14, 381– subfamilies and circumscription of tribes in the family Hesperiidae 390. (Lepidoptera: Hesperioidea). Cladistics 24, 642–676. Mitchell, A., Mitter, C., Regier, J.C., 2000. More taxa or more characters revisited: Warren, A.D., Ogawa, J.R., Brower, A.V., 2009. Revised classification of the family combining data from nuclear protein-encoding genes for phylogenetic analyses Hesperiidae (Lepidoptera: Hesperioidea) based on combined molecular and of Noctuoidea (Insecta: Lepidoptera). Syst. Biol. 49, 202–224. morphological data. Syst. Entomol. 34, 467–523. Mitchell, A., Mitter, C., Regier, J.C., 2006. Systematics and evolution of the cutworm Weller, S.J., Pashley, D.P., Martin, J.A., Constable, J.L., 1994. Phylogeny of noctuoid moths (Lepidoptera: Noctuidae): evidence from two protein-coding nuclear moths and the utility of combining independent nuclear and mitochondrial genes. Syst. Entomol. 31, 21–46. genes. Syst. Biol. 43, 194–211. Mutanen, M., Wahlberg, N., Kaila, L., 2010. Comprehensive gene and taxon coverage Wheat, C.W., Wahlberg, N., 2013. Phylogenomic insights into the Cambrian elucidates radiation patterns in moths and butterflies. Proc. R. Soc. Biol. Sci. Ser. explosion, the colonization of land and the evolution of flight in Arthropoda. B 277, 2839–2848. Syst. Biol. 62, 93–109. Nieukerken, E., Kaila, L., Kitching, I., Kristensen, N., Lees, D., Minet, J., Mitter, C., Yang, L., Wei, Z.J., Hong, G.Y., Jiang, S.T., Wen, L.P., 2009. The complete nucleotide Mutanen, M., Regier, J., Simonsen, T., 2011. Order Lepidoptera Linnaeus, 1758. sequence of the mitochondrial genome of Phthonandria atrilineata (Lepidoptera: Zootaxa 3148, 212–221. Geometridae). Mol. Biol. Rep. 36, 1441–1449. Nylander, J., 2004. MrModeltest v2. Program distributed by the author, vol. 2. Yang, X.S., Xue, D.Y., Han, H.X., 2013. The complete mitochondrial genome of Biston Evolutionary Biology Centre, Uppsala University. panterinaria (Lepidoptera: Geometridae), with phylogenetic utility of Rambaut, A., Drummond, A., 2007. Tracer v1. 5.