We combined the original GWAS signal, risk haplotype, binding affinity significance and conservation information to prioritize the leading variants (L) detected by GWAS. For each L, we first find all variants (V) with GWAS3D signal and in LD with L (r2 > user selected cutoff). We then calculated V’s phenotypically associated effect (PGWAS) by dividing the GWAS P-value of L with the r2 between V and L in the user-specified populations. We then selected the most significant P-value of the LOD related to a specific TF motif (PBDA) to represent the binding affinity effect of the variant. We further mapped the variant to genomic evolutionary rate profiling++ (37) constrained elements and calculated the corresponding conservation P-value (PCONS). Using the P-values of aforementioned three independent measurements (GWAS, binding affinity, conservation), we performed Fisher’s combined probability test to calculate a combined P-value, CP, for each V. We then assigned the most significant CP from all the variants Vs to the corresponding leading variant L. All the Ls are then re-ranked according to their new CP values, with special focus on their regulatory effects.
