We provide a simple tool for post processing to make the results more readable.
We may want to see the results when there is no population structure correction. In this case, we create an identity matrix as the kinship matrix. Since we are only interested in those SNPs significant in the previous analysis, we don't need to check all SNP pairs.
In order to do this, I made another version of Pylmm-epi-set, which is called “pylmm-epi-partial-bf”. This version takes the output of “pylmm-epi-set”, and it only computes the p-value of SNP pairs specified by the output of “pylmm-epi-set”.
Up to now, we have obtained the p-values of all signficant SNP pairs, with/without population structure corrections. We need further process to make these results readable.
Post Step One:
Generate the the report for each possible causal haplotype.
We select one representative SNP pair from each cluster, under each
Post Step Two:
Combine significant SNP regions from different causal haplotype.