Phylogenomics resolves major relationships and reveals significant diversification rate shifts in the evolution of silk moths and relatives

Background Silkmoths and their relatives constitute the ecologically and taxonomically diverse superfamily Bombycoidea, which includes some of the most charismatic species of Lepidoptera. Despite displaying spectacular forms and diverse ecological traits, relatively little attention has been given to understanding their evolution and drivers of their diversity. To begin to address this problem, we created a new Bombycoidea-specific Anchored Hybrid Enrichment (AHE) probe set and sampled up to 571 loci for 117 taxa across all major lineages of the Bombycoidea, with a newly developed DNA extraction protocol that allows Lepidoptera specimens to be readily sequenced from pinned natural history collections. Results The well-supported tree was overall consistent with prior morphological and molecular studies, although some taxa were misplaced. The bombycid Arotros Schaus was formally transferred to Apatelodidae. We identified important evolutionary patterns (e.g., morphology, biogeography, and differences in speciation and extinction), and our analysis of diversification rates highlights the stark increases that exist within the Sphingidae (hawkmoths) and Saturniidae (wild silkmoths). Conclusions Our study establishes a backbone for future evolutionary, comparative, and taxonomic studies of Bombycoidea. We postulate that the rate shifts identified are due to the well-documented bat-moth “arms race”. Our research highlights the flexibility of AHE to generate genomic data from a wide range of museum specimens, both age and preservation method, and will allow researchers to tap into the wealth of biological data residing in natural history collections around the globe. Electronic supplementary material The online version of this article (10.1186/s12862-019-1505-1) contains supplementary material, which is available to authorized users.

for particularly large or egg-containing abdomens. Both tissue size and type played a significant role in determining which samples would be best for sequencing. For example, the bigger the abdomen, the more lysis buffer is needed to fully immerse during the digestion incubation step.
However, when the abdomen was large and hollow, more proteinase K was used to ensure that protein digestion could take place and not simply cell lysis. Because lipid rich tissue can interfere with the tissue digestion, as well as change the chemistry of the DNA isolation buffers, specimens that appeared "greasy" (i.e., high fat content) were deemed not ideal for this extraction method.
After incubation, samples were spun down briefly using a mini centrifuge to collect any droplets of solution. Abdomens were removed from the extraction buffer using sterile forceps, that had been first cleaned with a 10% bleach solution and then with 100% ethanol between each sample. Abdomens were gently rinsed with ≥95% Ethanol and stored at -20°C in ≥95% Ethanol to await genitalia dissection. This modified protocol allowed DNA extraction without compromising genitalia preparation for future taxonomic or morphological work. For samples containing 300 µL of buffer solution, 100 µL of Chloroform (Sigma Aldrich Cat# C2432-500) was added. For samples utilizing 500 µL of buffer solution, 200 µL of Chloroform was used. Samples were then vortexed and centrifuged for 10 minutes at 14,000 RPM. Supernatant was pipetted off into a new, clean, labeled 1.5 mL microcentrifuge tube. Excess chloroform and sample detritus were discarded into the proper hazardous waste container. For samples containing 300 µL of supernatant, 25 µL of DNA stripping buffer (OmniPrep Kit) was added. For samples consisting of 500 µL of supernatant, 50 µL of DNA stripping buffer was used. Samples were then vortexed and spun down using a mini microcentrifuge, followed by incubation in the dry air oven at 56°C for 10 minutes.
Once incubation was complete, samples were left to cool to room temperature for approximately 10 minutes. For samples containing 300 µL of supernatant, 50 µL of precipitation buffer (OmniPrep Kit) was added. For samples with 500 µL of supernatant 100 µL of precipitation buffer was used. During this step, all samples also received 5 µL of mussel glycogen (OmniPrep Kit). Samples were then vortexed to mix and centrifuged for 20 minutes at 14,000 RPM. The supernatant was pipetted off into a new, sterile, labelled 1.5 mL microcentrifuge tube, leaving a salt pellet in the original tube that was discarded. For samples containing 300 µL of supernatant, 250 µL Isopropanol (Fisher Scientific Cat# BP2618-500) was added to each sample. For samples of 500 µL of supernatant, 500 µL of Isopropanol was used.
Samples were vortexed and spun down using a mini centrifuge, then incubated at 4°C for 30 minutes. This step is a "stop step", allowing for the samples to be incubated in the freezer overnight, if necessary.
After incubation, samples were centrifuged for 10 minutes at 14,000 RPM. Excess Isopropanol was removed via pipette into the proper waste container. Once the Isopropanol was removed, 350 µL of 80% Ethanol was added to each sample, followed by the samples being centrifuged for 10 minutes at 14,000 RPM. Ethanol was poured off and samples were spun down using the mini centrifuge so that the remaining ethanol could be pipetted off. Sample tubes were then left open, but covered with a Kimwipe™ for 5-10 minutes to allow the remaining ethanol to evaporate. 50 µL of TE Buffer (OmniPrep Kit) and 0.5 µL of RNAse (OmniPrep Kit) were then added to each sample. Samples were vortexed and spun down using the mini centrifuge and stored at 10°C. DNA quality and quantity were evaluated through agarose gel electrophoresis and fluorometry using a Qubit 2.0 (Invitrogen, Thermo Fisher Scientific) to measure the relative DNA concentrations of each sample. Although yielding high amounts of gDNA, extracts from some museum specimens were too fragmented to be utilized for AHE sequencing. High quality gDNA was determined to be ≥2000bp; ≥1000 bp was deemed "good"; ~300 bp may work, but our experience is that not all of the loci will be recoverable; low quality gDNA is classified as ≤300 bp and will not capture or sequence well. DNA extracts were stored, as above, in the MGCL molecular collection.

Sphingidae
Historically, the Sphingidae have been considered a "highly derived" group (Brock [2], Holloway [3], Lemaire & Minet [4]), which, because of their divergent morphology, have sometimes been assigned to their own monotypic superfamily, Sphingoidea, rather than being included within Bombycoidea, where all recent morphological and molecular phylogenetic studies have now placed them. However, their position within the superfamily, with regard to families Saturniidae and Bombycidae (in the current sense, see below), has been uncertain. The current study confirms a sister relationship between Sphingidae and Saturniidae, with Bombycidae as their sister, in the SBS group.
However, the relationships among the sphingid genera and their assignment to suprageneric taxa have yet to be fully elucidated. The history of the classification of Sphingidae was summarized by Kitching & Cadiou [5], who then provided a slightly updated classification in which they recognized three subfamilies: Smerinthinae, containing three tribes, Smerinthini, Sphingulini, and Ambulycini; Sphinginae, with two tribes, Sphingini and Acherontiini; and Macroglossinae, with three tribes, Dilophonotini, Philampelini, and Macroglossini. The first of these was then divided into two subtribes, Dilophonotina and Hemarina, and the last also into two subtribes, Macroglossina and Choerocampina. Genera in each tribe and subtribe were separated into unnamed subgroups, some of which were indicated as possibly being monophyletic pending further study. Changes from the previous classification were essentially limited to moving Sphingulini from Sphinginae to Smerinthinae (based on larval similarities) and adjusting the placement of some genera.
The classification of Kitching & Cadiou [5] was, however, based on intuitive reasoning rather than phylogenetic analysis. The first phylogenetic analysis was that of Regier et al. [6], based on two nuclear gene sequences. Although the sampling was very limited (14 genera), they recovered strong support for a sister relationship between Smerinthinae and Sphinginae and evidence that the macroglossine tribe Dilophonotini was paraphyletic relative to Philampelini + Macroglossini. Kawahara et al. [7] expanded the sampling to five genes and 50 genera and found numerous novel relationship patterns. Again, Smerinthinae and Sphinginae were recovered as sisters, but only if Langia Moore was excluded from the former; this genus was found to be sister to Smerinthinae + Sphinginae. The smerinthine tribe Ambulycini was also monophyletic within a paraphyletic Smerinthini. Tribe Sphingulini was placed with a monophyletic Sphinginae (contra [5] but agreeing with pre-2000 classifications) though paraphyletic, with the Australian genera branching off first, followed by the Asian taxa. Tribe Acherontiini was also recovered as monophyletic, but nested deep within the Sphinginae. Druce, and, based on intuitive reasoning, proposed that their relationships within the subfamily were: Pentateuchini + (Monardini + (Hopliocnema genus-group + (Sphingulini + Sphingini))).
A new formal higher taxonomy for Sphingidae based on these results and previously implemented in part in the online Sphingidae Taxonomic Inventory ( [11]; accessed on 5 November, 2018) was published by Kitching et al. [12], in which three subfamilies, Langiinae, Smerinthinae and Sphinginae were recognized. Tribal and subtribal names were resurrected for many of the clades recovered by Kawahara et al. [7] and informal genus-groups proposed where no prior family-rank names were available.
The present study includes only 22 genera of Sphingidae but has nevertheless revealed several previously unknown relationships. The most unexpected was a strongly supported sister paring between Langia and Pentateucha. This suggests that Pentateuchini should be included as a tribe in Langiinae not Sphinginae. Additionally, Sphinginae could be rendered monophyletic also not unexpected, although more representatives of this large genus will need to be studied before a formal subdivision of the genus is made.

Saturniidae
The Saturniidae are the phylogenetically most well-studied family of bombycoid moths.
Each of the subfamilies and tribes (including those that were previously treated as separate families: Cercophaninae and Oxyteninae) have long been recognized to be monophyletic on the basis of strong morphological evidence ( [4], [13], [14]). However, the relationships among these clades were poorly understood and often conflicting, even within single works [15]. More recently, Saturniidae have been the focus of several molecular phylogenetic studies ( [16], [17], [18]), or included in broader Bombycoidea or Lepidoptera phylogenetic studies with strong taxon sampling ( [15], [19]). Unsurprisingly, the results of these studies have not found any novel subfamily or tribal groups, but they have deepened our understanding of the relationships among these taxa. All recent major molecular phylogenies that included Oxyteninae and Cercophaninae have found these two subfamilies to be sister to the remaining Saturniidae lineages, with Oxyteninae being sister to Cercophaninae + the remainder of the family. However, the relationships among the remaining subfamilies, Agliinae, Arsenurinae, Ceratocampinae, Hemileucinae, Salassinae, and Saturniinae, are less clear. Arsenurinae + (Ceratocampinae + Hemileucinae) ( [16], [17], [19]), a pattern supported by our Pr+Fl analysis; or in a sister relationship with Arsenurinae based on larval morphology (shared possession of tree-like scoli, which could be interpreted as symplesiomorphic). However, our AA analysis unexpectedly suggested that Agliinae were sister to Salassinae + Saturniinae, thereby increasing ambiguity in our understanding of the relationships of this small Old World subfamily.
Yet the most problematic suprageneric relationships are in Saturniinae, due to the historically uncertain composition of and relationships among the African subgroups: "Bunaeinae/-ini", "Microgoninae/-ini" and "Urotini". Regier et al. [16] showed a monophyletic Bunaeini and a paraphyletic Urotini. Barber et al. [17] included representatives of all the major Saturniinae lineages but failed to resolve their relationships fully. Rubin and Hamilton et al. [18] densely sampled Saturniinae, using hundreds of AHE loci, to provide the clearest picture so far of the relationships within this subfamily. These authors, like [16] and [17], did not recognize "Bunaeinae" as a subfamily, but instead included it as a tribe within subfamily Saturniinae, together with tribes Saturniini, Attacini, and the less clearly delimited Urotini and Micragonini.
Although the monophyly of and sister relationship between Attacini and Saturniini is supported by all these studies, as well as our own, the relationships among the other predominantly African tribes remain highly contentious. Our sampling of Saturniinae was not dense enough to disambiguate the relationships within this subfamily fully. The recently published results of Rubin and Hamilton et al. [18] did provide further insights, but unfortunately they did not discuss the systematics of the subfamily, despite offering the best published dataset so far.
Their study showed that the genus Usta Wallengren, previously considered a member of tribe Urotini [12], is actually sister to the tribe Micragonini. The remaining Urotini genera form a paraphyletic grade, nested within a monophyletic tribe Bunaeini.
Lastly, our results provide strong support for the sister-group relationship of Saturniidae and Sphingidae, corroborating an unpublished synapomorphy in adult antennal morphology (A. Zwick, pers. obs.).

Apatelodidae/Bombycidae/Endromidae/Phiditiidae
Historically, most genera that are now dispersed among the four families Bombycidae, Apatelodidae, Phiditiidae, and Endromidae were consolidated into a broader concept of Bombycidae ( [4], [14]). The exceptions were Endromis Ochsenheimer and Dalailama Staudinger, which comprised Endromidae, and Mirina Staudinger, the sole representative of Mirinidae. This view changed radically with the advent of molecular systematics [19], when the genera Oberthueria Kirby ("Oberthueriinae") and Prismosticta Butler ("Prismostictinae"), together with Mirina, were transferred into a broader concept of Endromidae. At the same time, Epiinae were upgraded from tribal to subfamily status in Bombycidae, and Apatelodidae and Phiditiidae were accorded family status. Our results strongly support the interfamilial findings of Zwick et al. [19], Wang et al. [20], and Lin et al. [21]. There is a clear reciprocally Following Zwick [22] and Zwick et al. [19], we treat Apatelodidae as a well-supported family, well differentiated from Bombycidae, with which they were long considered closely related ( [4], [14], [23]). Our study is the first to include more than two apatelodid genera, which confirms the monophyly of the family and provides a much-needed insight into the intrafamilial relationships of this poorly studied group. All our analyses recover the Apatelodidae genera as a grade within a well-supported monophyletic group, without clearly defined subclades, suggesting subfamilies are not warranted in this family. However, we were unable to sequence Carnotena Walker, Thelosia Schaus and Thyrioclostera Draudt, and denser taxon sampling within Apatelodidae may reveal finer details of the intergeneric relationships.
Our results also show that the only sampled genus currently (sensu [12]) misplaced in a bombycoid family is Arotros Schaus. This genus has been included in Bombycidae since its original description. However, our molecular phylogenetics results show that this unique monotypic genus is actually deeply nested within Apatelodidae. Interestingly, Arotros displays a fascinating degree of convergent evolution in habitus with the bombycine genus Rotunda, to which it is unrelated, relatively speaking, within Bombycoidea (both Arotros and Rotunda are sampled here). Based on this robustly supported placement, we transfer Arotros to Apatelodidae and recommend further research into its morphology to reinforce its move from Bombycidae.

Carthaeidae
This enigmatic, monotypic Australian family has been difficult to place with molecular systematics. Regier et al. [24] recovered it as sister to either Anthelidae or Endromidae, depending on the analysis, whereas Zwick [22] recovered Carthaeidae in various uncertain topological placements. Zwick et al. [19] was the first to recover a better supported sister relationship between Carthaeidae and the small Neotropical Phiditiidae, which was also the most robustly supported placement in our research (Pr+Fl analysis). This placement however, is still inconclusive due to the recovery of a weakly supported sister relationship between Carthaeidae and Endromidae in our AA analysis.

Brahmaeidae
A close association between Brahmaeidae and Lemoniidae had long been postulated ([4], [14], [25], [26]), but it was not until Zwick [22] that the two families were combined based on strong phylogenetic evidence. Our results support this decision with genus Lemonia Hübner nested among typical brahmaeid genera. We were unable to sample the second "lemoniine" genus, Sabalia Walker, but Zwick et al. [19] found strong support for this genus as sister to

Lemonia.
Our study is the first to use molecular phylogenetics to address the position of the enigmatic African genus Spiramiopsis Hampson, which has been variously included in "Sabaliadae" (= "Lemoniidae"), Eupterotidae, Bombycidae and Brahmaeidae. However, our results fully support its inclusion in the last of these families, as sister to Dactyloceras Mell. The only brahmaeid genus not yet sampled for molecular phylogenetics is Calliprogonos Mell, and its phylogenetic placement within the family remains unknown.
Lemonia and Sabalia are superficially quite distinct both from each other and from the large, intricately patterned Brahmaea and Dactyloceras (and, to a lesser extent, the more simply patterned Calliprogonos and Spiramiopsis). Thus, the clade comprising Lemonia and Sabalia could conceivably be treated as a subfamily, Lemoniinae. However, the African Dactyloceras and Spiramiopsis, which are sister to Lemonia + Sabalia, if considering [19], and their collective sister genus, Brahmaea, would then also need to be treated as subfamilies (and no name exists for the former). Given that the position of Sabalia remains to be confirmed and that of Calliprogonos to be determined, we do not at this time formally assign names to clades within Brahmaeidae, though it would surely be of interest to determine potential synapomorphies for the Lemonia, Sabalia, Spiramiopsis and Dactyloceras clade relative to Brahmaea to see whether a subfamilial arrangement is indeed warranted.

Eupterotidae
The current suprageneric classification of Eupterotidae follows Forbes [26], as modified by Oberprieler et al. [27] and Nässig & Oberprieler [28], comprising four subfamilies and one informal genus group. The subfamily Striphnopteryginae was considered to be monophyletic, as was Janinae, provided that the rather divergent genera Hibrildes Druce and Tissanga Aurivillius, previously included in their own subfamilies [4], were included. The composition of Eupterotinae has varied over time depending on how the tribe Cotanini was treated. Minet [14] excluded the tribe from Eupterotinae, transferring it to subfamily Panacelinae, and was followed in this regard by Lemaire & Minet [4], Holloway et al. [29] and Zwick [22]. In contrast, Oberprieler et al. [27] and Nässig & Oberprieler [28]  Panacela Walker. Finally, the "Ganisa-group" was proposed as an informal group for those genera not assignable to any of the four subfamilies and as such is most probably paraphyletic, or even polyphyletic. However, this classification of Eupterotidae is not underpinned by any objective phylogenetic analysis. The only objective study of Eupterotidae to date, using molecular sequence data, is that of Zwick [22], but it included only four genera: Hoplojana However, this is perhaps not surprising as the subfamily has received little recent taxonomic attention and is probably just a dumping ground for superficially similar plain and nondescript species. If Cotana is accepted as a panaceline, then the large subfamily Eupterotinae is represented by only two taxa. An undetermined species of Eupterote Hübner is sister to Panacelinae, supporting a previous suggestion by Zwick [22] that the latter should be treated as a tribe (Panacelini) of the former. The position of Sphingognatha C. & R. Felder as sister to Striphnopteryginae + (Eupterotinae + Panacelinae) is somewhat unexpected as the genus has been synonymized with Eupterote [3]; although reinstated by Oberprieler et al. [27], on the basis of its different male genital morphology. Finally, and rather unexpectedly, the Ganisagroup was recovered as monophyletic, except for Neopreptos Draudt, which is placed as sister to the Eupterotinae/Striphnopteryginae/Panacelinae clade. This is one of only two New World eupterotid genera and is nested within a large group of Asian genera, which has very interesting biogeographical implications. Its position on a long branch also suggests that it may deserve its own subfamily (probably also including the other New World genus, Preptos Schaus).
Our results show several patterns of relationship that are incongruent with the current classification and suggest that the subfamily structure will need marked revision. However, the present study is still far from comprehensive and many other genera should be sampled and their relationships elucidated using phylogenetic analysis before any formal changes are proposed.
More complete sampling of genera may help to resolve this ambiguity, including especially, in the present context, the second anthelid subfamily, Munychryiinae.
The only reasonably comprehensive phylogenetic treatment of Anthelidae genera to date [22] found that the genus Anthela Walker was polyphyletic with regard to several of the other sampled antheline genera, a result also found in the present study. However, a further sampling within this genus (including the type species, Anthela ferruginosa Walker) will be necessary before a reclassification of the family can be proposed.
List of the loci names in the modified Bombycoidea-specific AHE probe set BOM1. The first column lists the name of the locus; the second column explains what the locus was named in the LEP1 kit or the name of the "legacy" or vision-related loci; the third column lists the length (bp) of each locus.
List of the total taxa included in the phylogenetic analyses. Specimen names refer to tip names, and include the specimen accession number and preliminary taxonomy. These specimen names can be used to identify a specimen's specific molecular data in the supplemental data. Other pieces of information include: taxonomy (family, subfamily, tribe, genus, and species -per Kitching et al. [12]); storage method of the tissues; whether the data came from an AHE probe set, transcriptome, or genome; and collecting date.
Maximum likelihood tree of Bombycoidea, based on 650 AHE loci. All nodes are supported by ≥95% bootstrap values unless otherwise noted. Major taxonomic groups such as families, subfamilies, and tribes are labeled. Red circles at the tips correspond to genera that are not monophyletic. Black boxes around the tips correspond to non-monophyletic subfamilies.
Interfamilial relationships of the CAPOPEM group. These relationships change depending on the data used (Pr+Fl or AA) to infer the phylogeny and the phylogenetic inference (supermatrix or ASTRAL). Values at the nodes of the Pr+Fl or AA trees indicate SH-aLRT/UFBS support, or ASV for the ASTRAL analysis.
Interfamilial relationships of the Saturniidae. These relationships change depending on the data used (Pr+Fl or AA) to infer the phylogeny and the phylogenetic inference method (supermatrix or ASTRAL). Values at the nodes of the Pr+Fl or AA trees indicate SH-aLRT/UFBS support, or ASV for the ASTRAL analysis.
The four most common 95% credible set of rate shift configurations sampled with BAMM.
Branch color indicates the estimated diversification rate, with warmer colors representing lineages with higher rates. Major taxonomic groups with shifts are labeled. Photographs correspond to the major lineages (i.e., the SBS group) with diversification shifts in the phylogeny.
The AA phylogeny.
This information as well as additional Supplemental Information can also be found on Dryad (will update upon acceptance with the unique persistent identifier and hyperlink to datasets in http:// format).