Hi, I have imputation results contain over 85 million SNPs for 21 samples. Over 64% of the SNPs have DR2 of zero and they are also monomorphic reference (AC == 0). To further validate the results, we calculate the Impute2-like INFO score using the bcftools plugin (impute-info). Surprisingly majority of the SNPs with DR2 zero have INFO score equal 1. Hence, we're not sure if we should exclude the SNPs with DR2 zero. Are there other metrics we should include in imputation QC?