Alzheimer's disease (AD) is the most common type of dementia in the elderly and has multiple causes. The amyloid precursor protein (APP), presenilin1 (PSEN1), and presenilin2 (PSEN2) genes were identified as causative genes in a small number of families with autosomal dominant early-onset forms of AD (ADEOAD). However, many AD cases are sporadic and the late-onset type, which develops after 65 years of age. The apolipoprotein gene (APOE) is the strongest risk gene for sporadic and familial late-onset AD (LOAD) regardless of ethnicity. Since about half of the patients with LOAD do not have the risk allele (ε4) of APOE and λs (=5) is larger than λsAPOE (=2.5), risk genes other than APOE are expected to be involved in LOAD. Based on the common disease-common variants hypothesis, genome-wide association studies (GWAS) were performed tenaciously to identify genes related to AD. However, even large-scale GWAS with relatively high frequency of single nucleotide polymorphisms did not reveal any additional risk genes with nearly equal power of APOE. Recently, individual whole genome or whole exon sequencing data obtained using next-generation sequencers have become available. Mutations in the causative genes of ADEOAD have been also observed in both familial LOAD and sporadic EOAD. Furthermore, new causative genes have been identifies in some families by whole genome or exome analyses. Considering the new common disease-rare variants hypothesis, personal genome sequence analysis is a potential strategy for identifying AD risk or protective genes.