Demographic scenarios used to test the performance of V and D matrices.

<p>Each simulation consists of 100,000 randomly ascertained SNPs (see <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1003480#pcbi.1003480.s003" target="_blank">Figure S3</a> for simulations with 10,000 SNPs) simulated with ms software in 25 populations (10 diploid individuals in each population). A) A.1) 2-D stepping stone demographic model implemented in the simulations. Each population exchanges a fraction of m migrants with the neighbour populations each generation. A.2) Amount of variation explained among populations by using either the D or V matrix in simulated against the inverse of the scaled amount of migrants by generation <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1003480#pcbi.1003480-Barton1" target="_blank">[57]</a>. A.3) Percentage of best genetic-matching partners (best overall match (BOM)) in the same (sub)population depending on whether the matrix V or D is used. The red dashed line indicates the simulation when the BOM computed from V matrix > BOM computed with D matrix. B) B.1) A sequential split demographic scenario. Each t generations the youngest population (at the right of the plot) splits into two. One remains in the same place and the new one moves to a new position at the right, decreasing its effective population size proportionally to the number of already conducted sequential splits. B.2) Percentage of SS(AP) respect to SS(T) by using either the D or the V matrix against the scaled time of split by Ne. B.3) Percentage of best genetic-matching partners (best overall match (BOM)) in the same (sub)population depending on whether the matrix V or D is used. The red dashed line indicates the simulation when the BOM computed from V matrix > BOM computed with D matrix.