Phylogenetics of the bat genus Eptesicus using UCEs. https://doi.org/10.1016/j.ympev.2022.107582 The bat genus Eptesicus is distributed globally and yet the phylogenetic relationships among species within this genus, especially in the New World, is not clear. Interestingly, molecular studies found that the New World Eptesicus are more closely related to Histiotus, another morphological genus having enlarged ears, than to the Old World Eptesicus. To better estimate the phylogeny of the New World Eptesicus and Histiotus species, we collected samples across taxonomic and geographic scales and genomic markers of thousands of ultra-conserved elements (UCEs). We further reconstructed ancestral distribution and tested hypotheses about the initial dispersal route (on-land versus trans-marine) of Eptesicus bats to colonize the New World. I sincerely thank all the museums and researchers who kindly provided samples to support this study!
My visit to the Smithsonian National Museum of Natural History.
Range-wide phylogeography of the big brown bat (Eptesicus fuscus) using RADseq. https://doi.org/10.1111/jbi.14362 The big brown bat (Eptesicus fuscus) is a common house bat in North America, distributed from southern Canada to norther South America including most of the Caribbean Islands. The previous study showed mitochondrial divergence but a lack of nuclear structure across the distribution range. To test whether gene flow has eroded the historical divergence (such as due to glacial isolation) in the nuclear genome, we collected range-wide samples and genome wide markers using the restriction site-associated DNA sequencing (RADseq). Analyses of population genetics, phylogeography, and demography supported historical isolation followed by secondary gene flow, showing signals of fine-scale nuclear divergence despite effects of on-going gene flow. In addition, species distribution modeling predicted further northward range shifts of this species under the future climate change.
Nonrandom missing data can bias the population structure indicated by Principal Component Analysis. https://doi.org/10.1111/1755-0998.13498 The Principal Component Analysis (PCA) requires no missing data in the input, which is not available in most non-model studies where sample qualities and quantities vary a lot. Therefore, the input missing data are imputed by the mean values (default setting in the R package adegenet). We show that the mean imputation would drag samples with relatively higher proportions of missing data to the origin of the PCA plot, thus potentially lead to misinterpretation of the illustrated population structure. We recommend plotting PCA with color-coded per-sample missing values to detect this potential bias.