Genvectors: exploring patterns and unveiling processes driving the geographic distribution of genetic lineages

Abstract

Sets of local populations show different degrees of gene flow due to dispersal barriers and environmental constraints, which renders genetic composition gradients among populations (genetic turnover). Unveiling biogeographic correlates of genetic turnover is paramount for phylogeography. While some processes (genetic drift, secondary contact) may erase the historical track of genetic turnover, vicariance or ancient dispersal likely leads to genetic divergence among populations. Yet available analyses do not permit direct inference about mechanisms driving genetic turnover. We propose a novel analytical approach called genvector analysis, which fulfills this gap by decomposing genetic compositional dissimilarities between populations based on either haplotypes or other genetic data into genetic eigenvectors. Such procedure allows exploring genetic turnover among sets of local populations, and analyzing their biogeographic correlates based on null model tests. We evaluate the statistical performance of the method on simulated datasets. We also analyzed biogeographic correlates of genetic turnover of Akodon cursor in the Brazilian Atlantic Forest. Results revealed that genvector analysis is robust to discriminate biogeographic drivers of genetic turnover. For Akodon cursor analysis, we observed that while for the entire species, all predictors considered (except for elevation) explained genetic turnover, within phylogroups some factors varied their importance. Genvector analysis was demonstrated to be useful for several different purposes in phylogeography, and complementary to classic analytical tools widely used by phylogeographers, such as AMOVA or DAPC. The role of ancient versus recent biogeographic events, the relationship between morphological divergence or abiotic variables and genetic turnover can easily be investigated using genvector analysis.