1. ¹Laboratory for Molecular Psychiatry, RIKEN Brain Science Institute, Saitama, Japan
2. ²Division of Population Genetics, National Institute of Genetics, Shizuoka, Japan
3. ³CREST, Japan Science and Technology Agency, Tokyo, Japan

Correspondence: Dr T Yoshikawa, Laboratory for Molecular Psychiatry, RIKEN Brain Science Institute, 2-1 Hirosawa, Wako-city, Saitama 351-0198, Japan. Tel: +81 48 467 5968; Fax: +81 48 467 7462; E-mail: takeo@brain.riken.jp

⁴The first two authors equally contributed to this work.

Received 5 August 2008; Revised 6 October 2008; Accepted 20 November 2008; Published online 17 December 2008.

Abstract

Associations have been reported between the variable number of tandem repeat (VNTR) polymorphisms in the exon 3 of dopamine D4 receptor gene gene and multiple psychiatric illnesses/traits. We examined the distribution of VNTR alleles of different length in a Japanese cohort and found that, as reported earlier, the size of allele '7R' was much rarer (0.5%) in Japanese than in Caucasian populations (20%). This presents a challenge to an earlier proposed hypothesis that positive selection favoring the allele 7R has contributed to its high frequency. To further address the issue of selection, we carried out sequencing of the VNTR region not only from human but also from chimpanzee samples, and made inference on the ancestral repeat motif and haplotype by use of a phylogenetic analysis program. The most common 4R variant was considered to be the ancestral haplotype as earlier proposed. However, in a gene tree of VNTR constructed on the basis of this inferred ancestral haplotype, the allele 7R had five descendent haplotypes in relatively long lineage, where genetic drift can have major influence. We also tested this length polymorphism for association with schizophrenia, studying two Japanese sample sets (one with 570 cases and 570 controls, and the other with 124 pedigrees). No evidence of association between the allele 7R and schizophrenia was found in any of the two data sets. Collectively, this study suggests that the VNTR variation does not have an effect large enough to cause either selection or a detectable association with schizophrenia in a study of samples of moderate size.