|| Checking for direct PDF access through Ovid
To search for schizophrenia susceptibility loci, we carried out a case–control study using 28601 microsatellite markers distributed across the entire genome.To control the highly multiple testing, we designed three sequential steps of screening using three independent sets of pooled samples, followed by the confirmatory step using an independent sample set (>2200 case–control pairs).The first screening using pooled samples of 157 case–control pairs showed 2966 markers to be significantly associated with the disorder (P<0.05). After the second and the third screening steps using pooled samples of 150 pairs each, 374 markers remained significantly associated with the disorder. We individually genotyped all screening samples using a total of 1536 tag single nucleotide polymorphisms (SNPs) located in the vicinity of ∼200 kb from the 59 positive microsatellite markers. Of the 167 SNPs that replicated the significance, we selected 31 SNPs on the basis of the levels of P values for the confirmatory association test using an independent-sample set. The best association signal was observed in rs13404754, located in the upstream region of SLC23A3. We genotyped six additional SNPs in the vicinity of rs13404754. Significant associations were observed in rs13404754, rs6436122, and rs1043160 in the cumulative samples (2617 cases and 2698 controls) (P=0.005, 0.035, and 0.011, respectively). These SNPs are located in the linkage disequilibrium block of 20 kb in size containing SLC23A3, CNPPD1, and FAM134A genes.Genome-wide association study using microsatellite markers suggested SLC23A3, CNPPD1, and FAM134A genes as candidates for schizophrenia susceptibility in the Japanese population.