addition to APOE exist. The objective of this study is to discover LOAD susceptibility genes other than the APOE. Methods: We conducted a three-stage genome-wide association study based on SNP in a total of 4,336 Japanese clinical subjects comprising 2,184 cases and 2,152 controls. In Stage 1, we genotyped 909,622 SNPs in 1,008 cases and 1,016 controls using Affymetrix GeneChip 6.0 microarrays. After stringent quality control (QC), 574,828 SNPs were analyzed in 951 cases and 894 controls. In Stage 2, we examined other sample sets consisting of 885 cases and 985 controls using Illumina Custom GoldenGate microarrays. After QC, 1,395 out of 1,418 SNPswere analyzed in 878 cases and 984 controls. In Stage 3, we genotyped 80 SNPs, comprising the high-ranked SNPs or SNPs of interest from Stage 2, in 1,977 cases and 2,128 controls by means of the TaqMan method. After QC, 80 SNPs were analyzed in 1,933 cases and 2,116 controls. Results: Multiple strong associations (Pallele < 1.7 3 10 ) for LOAD were observed in/around APOE region including BCL3, PVRL2, TOMM40,APOE,APOC1, and LRRC68 on chromosome 19. All association signals of these genes disappeared when subjects were stratified as to the APOE-e4 carrier status. In addition to these genes, we found several genes, albeit weak association, PAPOLOG (chr 2), ARHGEF3 (chr 3), EBF1 (chr 5), SORL1 (chr 11), FAM124A (chr 13) and GNAO1 (chr 16). To verify the association of these genes with LOAD in the Japanese population, we proceeded with replication studies in different ethnic cohorts, ADGC (11,840 cases and 10,931 controls) and Korean (333 cases and 1,128 controls).Conclusions: In addition to APOE, we found several candidate genes susceptibility to LOAD through a genome-wide association study.