Association study of nonsynonymous single nucleotide polymorphisms in schizophrenia

Genome-wide association studies using several hundred thousand anonymous markers present limited statistical power. Alternatively, association studies restricted to common nonsynonymous single nucleotide polymorphisms (nsSNPs) have the advantage of strongly reducing the multiple testing problem, while increasing the probability of testing functional single nucleotide polymorphisms (SNPs). We performed a case-control association study of common nsSNPs in Galician (northwest Spain) samples using the Affymetrix GeneChip Human 20k cSNP Kit, followed by a replication study of the more promising results. After quality control procedures, the discovery sample consisted of 5100 nsSNPs at minor allele frequency >5% analyzed in 476 schizophrenia patients and 447 control subjects. The replication sample consisted of 4069 cases and 15,128 control subjects of European origin. We also performed multilocus analysis, using aggregated scores of nsSNPs at liberal significance thresholds and cross-validation procedures. The 5 independent nsSNPs with false discovery rate q ≤.25, as well as 13 additional nsSNPs at p <.01 and located in functional candidate genes, were genotyped in the replication samples. One SNP, rs13107325, located at the metal ions transporter gene SLC39A8, reached significance in the combined sample after Bonferroni correction (trend test, p = 2.7 × 10 ^-6, allelic odds ratio = 1.32). This SNP presents minor allele frequency of 5% to 10% in many European populations but is rare outside Europe. We also confirmed the polygenic component of susceptibility. Taking into account that another metal ions transporter gene, SLC39A3, is associated to bipolar disorder, our findings reveal a role for brain metal homeostasis in psychosis.