BGEN format on UK Biobank data does not work properly

[jielab]

Hi, I use RVTESTS latest version to read the UK biobank BGEN file, a command like this: rvtest --inBgen ukb_imp_chr1_v2.bgen --inBgenSample ukb_imp.sample --pheno bmi.pheno.ped --pheno-name bmi --out chr1.rvtest --single score.

Then I compared the results from PLINK and from Ben Neale's group. I found that the results from RVTESTS is way off from the other two, while the other two are pretty much the same. So, I don't know if there is some issue, maybe you could test run it and find out. BTW, the sample order in my bmi.pheno.ped file is different from the ukb_imp.sample file. I hope that is no problem.

Attached below is a comparison on EAF, Beta, and P-value.

Best regards,
Jie

[zhanxw]

The sample order should not affect EAF calculation. Let me check again. Thanks very much.