Collection of polymorphic microsatellite markers:
We optimized pooled-DNA genotyping method which utilizes DNA mixture consisting of aliquot DNA from each sample to identify polymorphic microsatellite markers efficiently and to save costs. Polymorphism of all the microsatellite markers which were designed across the whole human genome were tested by using pooled-DNAs from 100 individuals of normal-healthy Japanese. At first, we collected 9,880 known microsatellite markers, which were polymorphic in Caucasian populations, and tested for their polymorphism in Japanese population: 9,099 of these known microsatellite were found to be polymorphic in Japanese. Then we designed 56,207 of new microsatellite markers across the whole human genome and screened: 30,950 markers displayed a high degree of genetic polymorphism. As a result, we obtained a total of 40,049 microsatellite markers which showed polymorphism in the Japanese population: the average resolution of these markers was one microsatellite every 75.0 kb, and the average heterozygosity and allele frequency were 0.69 and 7.0, respectively. Therefore, we succeeded in collection of microsatellite markers which show polymorphism in the Japanese population across the entire human genome with high density enough to be utilized for genome-wide case-control study.
Genetic homogeneity of Japanese population:
We classified 116 Japanese samples based on Y chromosomal lineage and investigated allele frequencies of 26 microsatellite markers on the autosomal chromosomes. Any markers did not show any significant difference in allele frequencies among these groups, suggesting that there are no hierarchical or stratified population structures in our Japanese samples. This result supports the possibility of application for genome-wide association study using polymorphic microsatellite markers.
Survey of genes associated with complex disorder in HLA region:
We performed case-control analyses on skin cancer, azoospermia, and rheumatoid arthritis as target diseases in the human leukocyte antigen (HLA) region by precise mapping of their candidate gene regions using microsatellite markers. There was a skin cancer susceptibility region located 180kb telomeric of the HLA-C gene, and also a 200kb candidate segment for azoospermia located near the HLA-DQ and -DR loci. These successful outcomes ensure the certainty of searching for susceptibility regions of common or complex diseases by genome-wide association analysis using polymorphic microsatellite markers. In addition, it was confirmed again that a set of our microsatellite markers have appropriate density because the observed ranges of the thus identified susceptible candidate regions which showed linkage disequilibrium(LD) with nearby microsatellite alleles were estimated to be 100-200kb.
Likewise, we found the candidate gene segment for rheumatoid arthritis which is located telomeric region of the HLA-class III region by using polymorphic microsatellite markers, then a susceptibility gene of this complex disorder being successfully mapped and identified by following SNP-based association analysis. The ranges of the LD region observed in SNPs were shorter than those of microsatellites. These results demonstrated that our approach of microsatellite-based genome-wide study is quite expecting to survey susceptible genes complex disorders: namely, mapping of candidate regions by genome-wide association analysis by using polymorphic microsatellite markers at first, followed by identification of disease responsible genes by association analysis with SNPs within candidate regions.
Genome-wide surveys of genes associated with complex disorders:
We proceed to mapping of susceptibility regions by genome-wide case-control studies for rheumatoid arthritis, psoriasis, hyper tension, and strong myopia. Forty-seven candidate regions which were associated with rheumatoid arthritis (RA) were identified : we applied case-control pooled-DNA genotyping method for the first, second, and third screening to RA, and then markers which showed statistically significant difference of allele frequencies between normal and patient populations throughout these three independent screenings were confirmed by individual genotyping analysis. Seven candidate regions which showed the top 7 strongest disease-association were analyzed for identification of RA susceptible genes with SNP-based association study, and seven susceptibility genes associated with rheumatoid arthritis were thus identified from each of seven candidate regions. Two among seven susceptible RA genes were known susceptibility genes and five represent new rheumatoid arthritis susceptibility genes. In this way, our microsatellite-based genome-wide association technique was established as a new and powerful method to survey disease-associated genes across the whole human genome to achieve our final goal of this project. We have been undertaking case-control analyses for other 20 common or complex diseases, such as psoriasis, hypertension, strong myopia etc by using our method, and could identify approximately 30 candidate regions for each disease. Although time-consuming and cost problems do not enable us to identify susceptibility genes of these diseases by SNPs genotyping of a lot of candidate regions narrowed down by genome-wide microsatellite-based association studies, we are planning to continue to and put considerable effort into these study for identification of as many as susceptible loci of common or complex diseases.