GWAS Analysis using Plink

Question

[123]

Hi everyone,

I have only one allele information for each SNP in my ped file. How can I do an association study in plink?

Here 0 = SNP1, 1 = SNP2 and so on.

How do I associate this phenotype to my SNPs?

Answer 1

[mari.lynne]

Are you sure PLINK is reading the file correctly and hasn't split the snp across as two columns, so col's 0 and 1 should be merged (Could have been a header issue?). Did you convert your data to PED at some point from vcf, as something could have gone wrong there.

Unless you expect your data to be haploid, which it might be? I don't think PLINK is well suited for such data, but if you do, I would maybe reformat your ped file by simple duplication so A becomes AA (this would require bash awk/sed commands). And then you then might want to check the REF/ALT alleles by generating a frequency file of your dataset

plink --file test --freq 

and then this should recode your ped file

plink --file test --set-hh-missing

There's a bit more info here, https://zzz.bwh.harvard.edu/plink/data.shtml, but this is just a very loose suggestion and I haven't tried it myself (nor do I think most people would have!)