IMPUTE2 limit to only one population of 1000G Reference

Question

[nhaus]

I am trying to impute genotypes for 92 individuals for which I have genotype information from a Illumina genotyping array. I know that all individuals come from Europe, so I think it would make sense to only include european individuals from the 1000g reference panel that I am using. However, going through the documentation of impute2, I was not able to find a option that would allow it. For shapeitv2 you can simply use the --include-grp flag, which does exactly what I want.

So I was wondering if there is a similar option for impute2 that I missed or if you have another idea how I could go ahead.

Cheers!

Answer 1

[LauferVA]

1. Impute2 is considered outdated.
2. 1000 genomes is no longer the largest available reference out there, and imputing against the largest, highest quality reference available is a big deal in terms of imputation accuracy of the output.
3. It is crucially important that you impute using a global reference. All the individuals are "European".. what does that mean exactly? Ethnically Caucasian? Are you 100.0000% certain that NONE of the 92 individuals have any genetic admixture in their background? I'd be surprised if that were true... if any of them are indeed admixed, you won't have an appropriate reference. Even if I knew that none were admixed definitively, I would STILL use a global background. It's been found over and over to perform better.