It seems likely that the genetic contribution of common variants to the risk of major psychiatric disorders is distributed across at least hundreds of variants of very small effect. Even very large studies are under-powered to detect a large number of such effects, without being overwhelmed by false positives. We think that variants that carry these modest risks are mostly in exons or in functional non-coding DNA. If we can combine several sources of information to identify functional DNA, we may be able to identify these risk SNPs better. We propose here an Empirical Bayes approach to combining genomic information with association information from Genome-wide Association Studies (GWAS), to identify individual SNPs and genes that are implicated in Schizophrenia.