Use the different male/female demographics of the X chromosome to infer sex of ancestors

From @hyanwong on December 5, 2016 9:0

Inferring the full ARG is comparable in the recent past to inferring a good proportion of the family tree, and we have family trees in the short term for a good number of species (some humans, obviously, but also some study populations of other species). The problem is that ARG inference has nothing to say about the identity of the reconstructed ancestors, because the lineage could have passed either through paternal or maternal lines (or a combination of both). However, one potential advantage of full ARG inference is that we might be able to use sex differences in inheritance to map inferred ancestors to real, known ancestors. This would provide a good method for testing the accuracy of the ARG inference in a large number of cases.

The most obvious difference to use is the difference in recombination / Ne of the X chromosome in males vs females. Another could be the difference in recombination rate between male and female gamete production (I think such a difference exists). There may be others we could think of too.

Copied from original issue: mcveanlab/treeseq-inference#2

hyanwong / treeseq-inference

Use the different male/female demographics of the X chromosome to infer sex of ancestors #2