Statistical phylogeography from individual, de novo genome assemblies

The use of short-read genomic data for inferring phylogeography in non-model organisms has been hindered by the difficulty of genome assembly. Here we develop a simple pipeline for generating large numbers of outgroup-rooted alignments of orthologous sequence blocks from short-read data suitable for existing multi-locus methods. We test our bioinformatics approach on minimal samples of three individ- ual genomes, one from each major glacial refugium in the Western Palearctic oak gallwasp B. pallida. The resulting data provide unprecedented resolution about recent, Pleistocene history. We use maximum likelihood to quantify the support for alternative models of population divergence with admixture. Our analysis reveals support for an unexpected history, dominated by recent and sudden admixture from the Middle East to the Iberia, most likely via North Africa during the last glacial period. Despite this major admixture event, we are able to reconstruct the underlying order and timing of refugial population diver- gence. We show that our inferences are robust to the length of sequence blocks, the mutation model and the sampling location of individuals within refugia. The combination of de novo genome assembly and numerical likelihood calculation provides a powerfull and robust population genomic framework for es- timating phylogeographic history that can be applied to any organism without the need for prior genetic resources.