search

johannessonlab / HetVPaper

Scripts and pipelines for the Ament et al. (2022) NEE paper on allorecognition genes and reproducive isolation
MIT License
0 stars 0 forks source link

LD filtering #2

Closed SLAment closed 3 years ago

SLAment commented 4 years ago

In HetVPaper/SNPpop/scripts/PCASNPpop_SNPRelate.R, Sergio pointed out that in the function snpgdsLDpruning I should be using the option slide.max.bp to sample down the SNPs used and make the windows more independent. Probably slide.max.bp = 5000, or something like that based on the LD decay curves. I had it before with the default, which is 500000!

SLAment commented 3 years ago

I now added snpgdsLDpruning = 10000 to match the window analyses and because the LD decay is very low in most chromosomes at that point. It also turned out I had a few errors with the PCoA but I think it should be fixed now.