PLINK analysis for case/control studies

Question

[kc]

Hi,

I have genotype data from Affymetrix. I want to look at my cases and controls using gPLINK to see what SNPs are present and I want to find which SNPs are unique in cases compared to controls for comparison. What kind of QC steps and other analysis I should do to get the results I am looking for?

[Sam]

So you are trying to see if a particular SNP is captured in the case but not in control? If that is the case, the plink (not gPlink) command should be something like

plink --bfile <binary ped file prefix> --test-missing --out <output prefix>

You will then have a file that tells you whether if a snp has higher missingness in case / control (something similar to the thing you want) If a SNP is "unique" to case / control, then you should definitely get it. If you don't have the binary files but only the ped files, the you can replace bfile with file.

[kc]

Thank you very much. This is helpful.

Answer 1

[Maxime Lamontagne]

How many samples and SNPs do you have?

You should read this article (Quality control procedures for genome-wide association studies). They are showing how to use PLINK to do your QC.

[kc]

I have very few samples ~63. The article is really useful, thank you very much.