I am receiving imputed data from UK Biobank to conduct a GWAS on. Previously I have carried out GWAS on genotype data, which I have QC'd for missingness per individual and per SNP, sex discrepancy, MAF filter < 0.05, heterozygosity rate, cryptic relatedness and population stratification.

However, from reading several papers about QC of imputed data, I do not carry out any of these QC checks on the imputed data. What QC procedures should I expect to carry out on UKBB data? Just to filter by info score?

Or would it be better to carry out the initial QC checks on the genotype data and then impute myself?

  • 2
    $\begingroup$ In general my advice for stuff which is as subjective as QCing is to find a recent paper in a high profile journal (Nature Genetics perhaps) which has done a GWAS on imputed genotypes and see what they did in terms of QCing. There would be no need to impute yourself though. $\endgroup$
    – user438383
    Sep 10 at 10:23
  • 1
    $\begingroup$ A lot of effort went into the imputation of UK biobank by people who have been developing tools for imputation for many years. I don't think you'll be able to do better unless you are such a methods developer yourself or have access to a really nice secret reference panel. disclaimer: I developed tools that were likely used in the imputation of the uk biobank data that you are working with. $\endgroup$
    – winni2k
    Sep 11 at 18:38
  • $\begingroup$ Thanks both for the comments, they were very reassuring! $\endgroup$
    – icedcoffee
    Sep 12 at 11:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.