SNP2HLA merge dosage files with overlaping samples

Question

[datanerd]

HI,

I did imputation with Hapmap and I have my dosage files. But the thing is I overlapped samples with different batches for running imputation (for e.g. 1 through 200 –one batch for imputation, then 150-300 next batch of imputation). Iam stuck as to how do I merge these into one file for association testing. I tried to find a way in plink but failed.

Please let me know if you are experienced with a way of doing it.

Thanks, Mamta

Answer 1

[Kevin Blighe]

It is more common to impute in 'chunks' of SNPs across the entire cohort. By dividing your dataset into sample groups, I feel that your imputation will be biased in some way.

I have posted some general guidance for pre-phasing and imputation, here:

• C: Phasing with SHAPEIT
• A: ERROR: You must specify a valid interval for imputation using the -int argument,

At the end, one converts the 'HAPS' chunks into VCFs, from where you can then piece everything back together nicely with BCFtools.

Kevin