Population networks from DNA sequences: methodological developments

Several classes of methods have been proposed for inferring the history of populations from genetic polymorphism data. As connectivity is a key factor to explain the structure of populations, several graph-based methods have been developed to this aim, using popu- lation genetics data. Here we propose an original method based on graphical models that uses DNA sequences to provide relationships between populations. We tested our method on various simulated data sets, describing typical demographic scenarios, for different param- eters values. We found that our method behaved noticeably well for realistic demographic evolutionary processes and recovered suitably the migration processes. Our method provides thus a complementary tool for investigating population history based on genetic material.