Hi all,

I am conducting GWAS and the problem is case and controls were genotyped using different chips and I got high inflation factor (lambda was almost 1.2).

All cases and some controls used chip A, and other three control data used chip B, C and D.

Because these data are from single population only, I performed imputation using same reference data.

What I could do was as below.

1. I selected variants with high info score from imputation (info>0.9)
2. I excluded variants which were significant between control population (control A vs. other controls)
3. I performed PCA, but it showed cases and control A are different population. Maybe it's because their phenotypes are different even though they're genotyped using same chips. Anyway when I used PC scores as covariates, inflation factor did not reduce.

Would you suggest any other methods I should try?

Thank you