New paper on Leptocircini (Papilionidae)

[adamcotton]

Just published online today:

Reboud, E.L., Nabholz, B., Chevalier, E., Lafon, B.J., Tilak, M., Mielke, C.G.C., Cotton, A.M & F.L. Condamine 2024. Clarifying the phylogeny and systematics of the recalcitrant tribe Leptocircini (Lepidoptera: Papilionidae) with whole-genome data. Systematic Entomology, 1–28.
Available open access from: https://doi.org/10.1111/syen.12661.

Abstract
Leptocircini is a dazzling tribe of Papilionidae, including dragontails, kite swallowtails and swordtails. This tribe is widely distributed, notably throughout the tropics of Africa, Southeast Asia and the Americas, making it a fascinating model in evolutionary biology. However, despite accounting for 25% of the global swallowtail butterfly diversity, Leptocircini have been surprisingly neglected in phylogenetic analyses. This has left unanswered questions about their taxonomy and systematics. Here, we present a new taxonomic working list for Leptocircini, featuring 162 valid species. Using a combination of long and short reads data, we produced five new reference genomes, and we generated highly covered and scaffolded whole genomes for 148 individuals to infer densely sampled phylogenetic hypotheses. Based on mitochondrial or thousands of nuclear genes and multiple phylogenetic approaches, a robust phylogenomic tree is recovered, representing approximately 90% of the known species, which allowed examination of several key phylogenetic hypotheses. We found the monotypic genus Protographium Munroe to be sister of genus Graphium Scopoli. Additionally, we found that subgenus Paranticopsis Wood-Mason and de Nicéville is nested within subgenus Pathysa Reakirt, which we found is likely attributed to an ancient gene flow. We therefore synonymize Paranticopsis, syn.rest. To keep a consistent approach to subgeneric classification across the tribe and family, we devided genus Eurytides Hübner into three subgenera: Mimoides Brown, Eurytides sensu stricto and Protesilaus Swainson. This led to several taxonomic implications: Asiographium Möhn, syn.rest., Boreographium Grishin, syn.n., Hyalaus Grishin, syn.n. and Neographium Möhn, syn.n. are synonymized with Eurytides (Mimoides); and Eurygraphium Möhn, syn.rest. is synonymized with Eurytides (Eurytides). Our analyses finally raised concerns about potential taxonomic inflation in two species-groups within Graphium and Eurytides (Protesilaus). This study illuminates the clade’s evolutionary history and paves the way for further research on this diverse group of charismatic butterflies.

Adam.

[Chuck]

Wow, thanks Adam. I can't wait to sit and read all that. Nice job!

[Cabintom]

No changes within Arisbe eh? Someone needs to take the plunge and invalidate G. abri.

[adamcotton]

Tom,

Unfortunately no samples of G. abri were available for sequencing, as well as several other suspect taxa in the subgenus. Check out the species information in the table, pages 7-8 for Arisbe.

One important finding was that Arisbe is a monophyletic group rather than the tailed or mimetic species being related to the Asian ones. This also implies that mimicry arose independently in the groups, and in Neotropical Eurytides species.

Adam.

[Chuck]

Adam can you explain Figure 6? I know what species groups are and see multiple dots, but don’t understand the axes and what I’m seeing. What are “deep” nodes as opposed to standard nodes? Does sarpedon have “deep” nodes? Thanks

[adamcotton]

This is way outside my field so I wrote to the lead author, Eliette Reboud, and she very kindly replied in some detail:

Here would be a detailed answer and explanation for Figure 6.
"
Figure 6 illustrates gene concordance factor (gCF) and site concordance factor (sCF) values. Each point corresponds to a node in the phylogeny, with the color indicating its ultrafast bootstrap (UFBS) value. This visualization represents phylogenetic information differently from traditional tree representations, but infact the informations are the same as those on the tree of Figure S12. This type of analyse and graph are inspired from a blog of Robb Lanfear, and you may find Robb Lanfear's blog helpful, as it explains these metrics and type of graph in greater detail (http://www.robertlanfear.com/blog/files ... ctors.html).

There are two main things to explain here.

- First, what are gCF and sCF? When constructing phylogenetic trees, metrics like bootstrap values estimate confidence in a node. However, bootstrap values often converge to 100 when analyzing many loci (see Minh, Hahn and Lanfear 2020 for more information about that), even if the evolutionary history is complex (e.g., due to incomplete lineage sorting or gene flow). Concordance factors (gCF and sCF) offer complementary insights:

Basically, the gCF counts how many individual gene trees (here the 1402 individual gene trees) support each node in the final tree, and how many support alternative topologies. So for each node, on its complete version, there are 4 values, the first one is the percentage of trees supporting the same topology as the final tree, the next two values are the percentage supporting 2 alternative topologies (quartet with the surrounding species), and finally the last value is the percentage of trees having a polytomy or other alternative topologies at that location. The gCF is a useful alternative to bootstrap because as you add more genes to your analysis, bootstrap scores tend to converge to 100, whereas the gCF may better reflects the complexity of a phylogenetic history and the uncertainties that we might always have.

sCF (site concordance factor): Similar to gCF but focuses on the proportion of informative sites in the alignment that support a given node. These metrics highlight whether gene trees and sequence sites converge on the same phylogenetic signal, providing a nuanced picture of evolutionary history.

- Second, what do "deeps nodes" mean? First, yes, they are standard nodes. In this context, “deep nodes” simply refer to "older/early" nodes in the phylogeny (or in the clade you are looking at) that separate major sub-goups within the phylogeny, as opposed to more recent nodes distinguishing extant species. For example, "Deep nodes in Graphium Pathysa/Paranticopsis" in Figures 6 refers to the "old" nodes that separate the differents species-group within this subgenus. Indeed, as stated in the rest of the article, something probably "complicated" happened in the early history of Pathysa and this causes some irresolution in phylogenetic analyses (and metrics) for the group.

So, does sarpedon have “deep” nodes?
Yes, theoritically sarpedon does have deep nodes, as it would simply refers to the "early/oldest" nodes within the sarpedon groups (even if here, these nodes are not so "old" as it is a quite recent radiation and not composed of many species). But importantly here for sarpedon and Figure 6, is says that recent nodes within the sarpedon group (look at Figures S12 and S13 to better see and situates these nodes) show low gCF values (~30%), meaning only about one-third of the gene trees agree on the consensus topology (the tree represented in Figure 5).


(extract of Figure S12 for the sarpedon group and sister clades. Values are Bootstrap/gCF/sCF).

Conclusions are not very easy to make based only on this information, but it shows that in this species-group, the 1402 genes have generally not enough information to clearly separate species, or are not informative enough to distinguish well between the “species” of this species group (whereas those same genes are informative in other part of the tree, e.g. the rest of the Graphium (Graphium) nodes do not have such low gCF values). It could mean that gene flows are still ongoing between these different “species”, and that they might not all deserve the status of species. Note that this result was found even with a quite conservative in the sampling of the group by only taking anthedon for the group anthedon/isander/choredon group from of Cotton et al. 2022. This means that further splitting the sarpedon group could result in never-resolved phylogenies.

In conclusion, there may be some structure in the sarpedon population, (and this is clearly what is expected in a population from different islands), but this population structure does not necessarily mean "complete" speciation. In biology, it has been shown that in some cases that populations can maintain strong structuring without leading to speciation, and phylogeny as measures of population structure does not appear to be necessarily good predictors of this speciation.
"

I am sorry it is a long response but it is actually not that easy to explain simply as it refers to a lot of different things. I hope it answers the question, though. I added the supplementary information (that is found online here https://resjournals.onlinelibrary.wiley ... pinfo.docx), where Figures S12-S13 can be found and help to visualize the metric on the tree.

All the best,
Eliette

[Chuck]

Now I understand that “deer in the headlights” look I get from naturalists I run into in the field who ask what I’m doing. I sort of understand all that, at a basic level. I will say it represents something other than I understood, so at least I’m better off.


Thanks Adam and please thank Eliette.

[adamcotton]

The print version of this paper has been published. Here is the citation pf the print version:

Reboud, E.L., Nabholz, B., Chevalier, E., Lafon, B.J., Tilak, M., Mielke, C.G.C. et al. (2025) Clarifying the phylogeny and systematics of the recalcitrant tribe Leptocircini (Lepidoptera: Papilionidae) with whole-genome data. Systematic Entomology, 50(2), 387–414.
Available from: https://doi.org/10.1111/syen.12661