OK, I had it with the validation site selector running for 60 hours (due to speed of genotype reading/parsing) in gsa3 only to fail in OnTraversalDone() because of some silly operator issue. Break up validation site selection process by chromosome, pick # of sites in each chromosome proportional to chr length, (taking care of roundoff issues to ensure precisely requested number of sites is kept), and then CombineVariants in the end. This also makes the selector run comfortably under 2Gb and thus can be easily LSF'ed

This commit is contained in:
Guillermo del Angel 2011-11-18 07:52:52 -05:00
parent 5c38a9cfd6
commit 99ed64933f

Diff Content Not Available