This article reviews recent genome-wide association studies (GWAS) of schizophrenia and considers future research directions. Until recently, genome-wide scans for disease risk variants were based on linkage analysis, an approach that can detect only those alleles that confer relatively large effects. The main alternative was the candidate gene association study, which, although better powered than linkage for weak genetic effects, is problematic for phenotypes where pathogenesis is largely unknown. However, systematic searches across the genome for risk alleles of small effect are now possible thanks to the development of array platforms allowing hundreds of thousands of single-nucleotide polymorphisms (SNPs) to be assayed, the genotypes of which are strongly correlated though linkage disequilibrium with a high proportion of the 10 million or so common SNPs in the human genome. Such GWAS have identified hundreds of genetic loci at which common variation confers susceptibility to many diseases and other polygenic traits,1,2 and this will increase as more phenotypes and larger samples are studied.