MAF in QC of genetic association study

Question

[GenMD]

Hello,

I'm using plink for QC. I have two questions re MAF:

1. My understanding is that plink (plink --file data --freq) will estimate the MAF based on the data and accordingly in the QC it will remove SNPs based on its own MAF estimates. Is this correct?
2. What is the standard for removing SNPs based on MAF threshold? Is it MAF estimates based on the data (as above)? or is it based on a general population e.g using 1000Genome data or similar?

Answer 1

[Maxime Lamontagne]

1. Plink will produce a file with the MAF for each SNP (plink.frq). You need to select which SNPs you want to remove.
2. In GWAS study: MAF < 1 or 5%.

[GenMD]

Thank you. Yes, I understand that. My question though how does plink calculate the MAF? is it based on my data or is it an estimate from the general population?

[Maxime Lamontagne]

Your data