NCI-CGR/GwasQcPipeline

dynamically adjust minimum MAF for HWE calculation and add option to set HWE p value threshold

Closed this issue · 2 comments

Currently, the MAF for HWE is set for all ancestries groups, but a larger sample size could allow lower MAF.
We would like:
Within each ancestry x platform group, flag SNPs with [MAF > sqrt(5/n) & HWE-p<p-val-threshold] for exclusion, where n=number of controls in that ancestry x platform group

@shukwong Do you have a reference for using sqrt(5/n) as the MAF cutoff that we can include in the docs or report?

No reference. But the idea is that we will expect >=5 alternative homozygous samples using sqrt(5/n) as a cutoff for MAF.