PT  - JOURNAL ARTICLE
AU  - Graham Jones
TI  - Divergence estimation in the presence of incomplete lineage sorting and migration
AID  - 10.1101/174342
DP  - 2017 Jan 01
TA  - bioRxiv
PG  - 174342
4099  - http://biorxiv.org/content/early/2017/08/09/174342.short
4100  - http://biorxiv.org/content/early/2017/08/09/174342.full
AB  - This paper focuses on the problem of estimating a species tree from multilocus data in the presence of incomplete lineage sorting (ILS) and migration. We develop a mathematical model similar to IMa2 (Hey, 2010) for the relevant evolutionary processes which allows both the the population size parameters and the migration rates between pairs of species trees to be integrated out. We then describe a BEAST2 package DENIM which based on this model, and which uses an approximation to sample from the posterior. The approximation is based on the assumption that migrations are rare, and it only samples from certain regions of the posterior which seem likely given this assumption. The method breaks down if there is a lot of migration. Using simulations, Leache et al. (2014) showed migration causes problems for species tree inference using the multispecies coalescent when migration is present but ignored. We re-analyze this simulated data to explore DENIM’s performance. We also re-analyze an empirical data set.