I am looking for a good way to compute similarity between pairs of genes for brain tissue. The purpose is to post-analyse GWAS data for the Alzheimer's disease. This is what I am currently thinking:
-- I "think" the best way is to find tissue specific eQTL SNPs for each genes which I guess I can find it from here (http://www.regulomedb.org/ ) and compute gene-gene similarity using gene-expression data. But I don't where to get this one. Is there any standard tissue specific eQTL for SNP ? Is there any standard gene co-expression matrix that I can download?
-- I tried to do that with GO terms but first of there are not tissue specific. Plus computing all Gene-Gen similarities are computationally very expensive. (BTW can I find it pre-computed somewhere?) Plus, mapping from SNPs to gene is just based on proximity (LD) which I guess too noisy for my application.
Any idea, suggestion?