Merge several association signals in a region and use as a phenotype

Question

[Floris Brenk]

Dear all,

I got some case/control association signals in one locus and now want to perform a bit more advanced eQTL analysis in this region. I already did a standard single variant eQTL analysis and here not much interesting came out. Now I want to merge the association signals (assuming they are somewhat independent) and create some kind of risk score per individual and use that value as input for the eQTL analysis. Is there a way that I can merge the beta's of the association signals into one score per individual? I have plink format genotypes and the variants risk alleles and beta's. Any ideas/methods?