Inherited or acquired dysfunction of the dopamine system is believed to underlie the core symptoms of schizophrenia, and there are some evidences that dopamine transporter activity may be altered in schizophrenic patients. Therefore, dopamine transporter gene (DAT1) has been traditionally considered a probable candidate gene for the association study of schizophrenia. Until now, association studies of the dopamine transporter gene (DAT1) with schizophrenia have yielded largely negative results. However, these results cannot be regarded as conclusive in that they were all obtained from just a single marker, that is, 3' untranslated region variable number of tandem repeat (VNTR). We have therefore tried to find other single nucleotide polymorphisms (SNPs) in DAT1 gene and to use them as additional markers for the association study of schizophrenia. Searching for the SNPs had been done with 50 Korean schizophrenic patients. DNA sequences encompassing the whole exon and flanking exon-intron junctions were amplified and searched for the presence of SNPs. Total of five SNPs were found. Among these, three SNPs (1215A>G, 1398C>T, IVS11+14G>A) as well as the 3' untranslated region VNTR were selected as the markers to be genotyped. The allelic and genotypic frequencies of these markers were determined in 252 schizophrenic patients and 271 controls and compared between them. The frequencies of algorithmically derived haplotypes were also compared. No evidence of association was found between any of these markers and schizophrenia. The result using haplotypes was also negative. However, when the patient subgroup with verified familial history and the subgroup with early age of onset were re-analyzed, weak trend of association between 1398C>T SNP marker with schizophrenia was found in both cases. In accordance with the previous literature, we could not find any evidence of association between DAT1 gene and schizophrenia. This result acquired more certainty because not only the VNTR but several SNPs present in DAT1 gene and newly constructed haplotypes were also used as additional markers. However, the finding of weak association between one of the SNP markers (1398C>T) and the specific subgroups of schizophrenia patients added further support to the importance of defining more homogenous subgroups in association studies.