A diversification relay race from Caribbean-Mesoamerica to the Andes: historical biogeography of Xylophanes hawkmoths

Abstract

The regions of the Andes and Caribbean-Mesoamerica are both hypothesized to be the cradle for many Neotropical lineages, but few studies have fully investigated the dynamics and interactions between Neotropical bioregions. The New World hawkmoth genus Xylophanes is the most taxonomically diverse genus in the Sphingidae, with the highest endemism and richness in the Andes and Caribbean-Mesoamerica. We integrated phylogenomic and DNA barcode data and generated the first time-calibrated tree for this genus, covering 93.8% of the species diversity. We used event-based likelihood ancestral area estimation and biogeographic stochastic mapping to examine the speciation and dispersal dynamics of Xylophanes across bioregions. We also used trait-dependent diversification models to compare speciation and extinction rates of lineages associated with different bioregions. Our results indicate that Xylophanes originated in Caribbean-Mesoamerica in the Late Miocene, and immediately diverged into five major clades. The current species diversity and distribution of Xylophanes can be explained by two consecutive phases. In the first phase, the highest Xylophanes speciation and emigration rates occurred in the Caribbean-Mesoamerica, and the highest immigration rates occurred in the Andes, whereas in the second phase the highest immigration rates were found in Amazonia, and the Andes had the highest speciation and emigration rates.

Evolution and Diversification Dynamics of Butterflies

“…We performed biogeographic stochastic mapping to examine in-situ speciation, immigration, and emigration between the 7-bioregions in BioGeoBEARS. We followed the protocol of Li et al 76 with 1,000 simulations with the DEC model. Relative mean dispersal rates between all permutations of bioregions were calculated and presented in Figure 3 (see also Data S5).…”

“…These mean dispersal rates represent dispersal of butterfly lineages throughout the entire evolutionary history of Papilionoidea, and thus cannot reveal changes in rates over time. In order to look at historical biogeography of butterflies during different epochs, rates along all possible inter-bioregion colonization rates were calculated at specific time intervals of 5 million years, following Li et al 76 (Table S27). These relative rates were then averaged to represent relevant geological time periods and presented in Figures S2-S4.…”

Large‐scale genomic data reveal the phylogeny and evolution of owlet moths (Noctuoidea)

“…We used SortaDate (Smith et al., 2018) to reduce the nucleotide alignment to a computationally tractable matrix (50 loci, 63.9 kb) using combined results of clock‐likeness (get_var_length.py) and least topological conflict with the species tree (get_bp_genetrees.py) follows Li et al. (2022a). The reduced concatenated data matrix was imported into BEAUTi (Drummond et al., 2012) with two different partitioning schemes: unpartitioned and partitioned by three codon positions.…”

“…We estimated divergence times within a Bayesian framework using BEAST v2.4.5 (Bouckaert et al, 2014) implemented on the highperformance computing cluster at the University of Florida, China Agricultural University, and the University of Memphis. We used Sor-taDate (Smith et al, 2018) to reduce the nucleotide alignment to a computationally tractable matrix (50 loci, 63.9 kb) using combined results of clock-likeness (get_var_length.py) and least topological conflict with the species tree (get_bp_genetrees.py) follows Li et al (2022a). The reduced concatenated data matrix was imported into BEAUTi (Drummond et al, 2012) with two different partitioning schemes: unpartitioned and partitioned by three codon positions.…”

Phylogeny and divergence time estimation of Io moths and relatives (Lepidoptera: Saturniidae:Automeris)

“…The topology generated from the PF dataset in IQTREE was used, given the higher bootstrap values when compared to the Probe dataset. However, when running BEAST using the PF dataset for dating, the resulting 95% confidence intervals were disproportionately small, indicating the high likelihood that the large dataset caused the underlying Bayesian analysis to become easily stuck at a local minima (Kawahara et al 2019;Rougerie et al 2022). Due to this, we ultimately chose to use the Probe dataset for dating.…”