One is the common variant hypothesis that attributes schizophrenia to the joint effect of multiple common genetic variants, with each contributing a small-to-modest risk of schizophrenia. The other is the rare mutation hypothesis, which proposes that schizophrenia is caused by rare mutations with a high clinical penetrance, and the pathogenic mutation is highly individualized in each affected patient or family. In this study, after re- sequenced the putative core promoter region, all the exons, and the 39UTR regions of the DLG4 gene in our patients, we did not find missense or frameshifting mutations in the DLG4 gene associated schizophrenia, indicating that exonic mutations in the DLG4 might be very rare in schizophrenia, and unlikely play a major role in the pathogenesis of schizophrenia. Nevertheless, we confirmed six polymorphic markers of the DLG4 gene that have been reported in the SNP database in our subjects, including rs2230178, rs6145976, rs2017365, rs739669, rs17203281, and rs13331. In silico analysis showed that different alleles of rs2230178, rs6145976, rs2017365, and rs739669 at the 59 end of the gene may have differential influences on expression of the DLG4 gene. The results of the reporter gene assay support the prediction from in silico analysis; the haplotypes MLN4924 had significant lower activity than the C-I-T-T. The rs17203281 at exon 12 is a synonymous mutation that does not alter the amino acid isoleucine at codon 432. Although several studies have indicated that synonymous mutations may not be always silent, we did not pursue the functional significance of this marker in this study, because this marker has been reported not to be associated with schizophrenia in a Han Taiwanese population in a previous study. The rs13331 at the 39UTR of DLG4 was also predicted to be functional, because it was located at the DICE site. The reporter gene assay showed C allele had a significantly higher Renilla activity than the T allele,Nilotinib suggesting the rs13331 may also play a role in regulating the expression of DLG4. In the genetic analysis, we found a different LD structure derived from 4 markers at the 59 end of the gene between the patient and control groups, suggesting that there might be different haplotype distributions between patient and control groups. The different LD might arise from the borderline association of the rs6145976 and the rs739669 with schizophrenia. Further haplotype-based association analysis showed a significant association of the haplotype C-D derived from two markers at the core promoter region with schizophrenia, and reporter gene activity assays showed that the haplotype C-D-C-C derived from rs2230178-rs6145976- rs2017365-rs739669 had significant lower promoter activity than that of the C-I-T-T, suggesting that C-D haplotype that is associated with schizophrenia may have reduced expression of the DLG4 gene in patients with schizophrenia. To our knowledge, there are only two genetic association studies of DLG4 with schizophrenia in the literature. Kawashima and colleagues chose six SNPs across the entire DLG4 gene and examined their association with schizophrenia in a Japanese sample ; they found no association of these SNPs and their derived haplotypes with schizophrenia; however, we found a borderline association of the rs13331 SNP with schizophrenia in our sample. In another study, Tsai and colleagues examined the association of two SNPs of the DLG4 gene with schizophrenia in a Han Taiwanese sample ; they found no association of these two SNPs with schizophrenia in their sample.