问题
I have a list of SNPs that I have already obtained their proxy SNPs with MAF and LD from 1000 genomes pilot 1. I want to know when everybody mentions about MAF matching, do they need to be exactly the same?
For example, the SNP of interest has MAF 0.35, if the proxy SNP has MAF 0.37, can it be used as a good proxy? Given LD of the two >0.8
Do have absolutely have to choose a proxy that has a MAF equals 0.35?
回答1:
LD as measured by R-sq is dependent on MAF ie high R-squared implies MAFs are similar. See:
https://www.ncbi.nlm.nih.gov/pubmed/18572214
So, very high R-squared usually means MAF between the SNPs are similar.
Also, selection of proxy SNPs is typically dependent on LD measure only. As mentioned MAF is "incorporated" into the calculation, so matching for it is not necessary. If you have multiple proxies with same LD, then choose one with most similar MAF.
来源：https://stackoverflow.com/questions/42399056/minor-allele-frequency-matching