@explodecomputer @mcgml @YiLiu6240 The current report will include the variance explained by the clumped SNPs (a sum of the r2 statistics for each individual SNP). Should we also add a QC metric reporting the variance explained by all SNPs regardless of their P value (the currently clumped SNPs must have P<5-e8)? Because the SNPs should be independent, I thought an easy way to do this would be to bin the genome into 1mb/10mb bins. We could add these bins to the GWAS file in Matt's pipeline. Then for the report, select the SNP with the smallest P value in each bin. Then calculate the r2 statistics for each of those SNPs and take the sum of those r2 values for the final report. What do you think? Would it be a useful metric?
@explodecomputer @mcgml @YiLiu6240 The current report will include the variance explained by the clumped SNPs (a sum of the r2 statistics for each individual SNP). Should we also add a QC metric reporting the variance explained by all SNPs regardless of their P value (the currently clumped SNPs must have P<5-e8)? Because the SNPs should be independent, I thought an easy way to do this would be to bin the genome into 1mb/10mb bins. We could add these bins to the GWAS file in Matt's pipeline. Then for the report, select the SNP with the smallest P value in each bin. Then calculate the r2 statistics for each of those SNPs and take the sum of those r2 values for the final report. What do you think? Would it be a useful metric?