68-1

5/98

psychiatric symptoms.Psychiatric disorders such as bipolar disorder and schizophrenia are highly heritable2). Observing that many patients have families or relatives with similar mental problems, clinicians have long noticed this fact. The high heritability of psychi-atric disorders is confirmed by the diagnostic concordance rate of monozygotic twins. Monozy-gotic twins have the same germline genetic infor-mation. The high rate of diagnostic concordance means that the contribution of genetic factors is highly relevant to psychiatric disorders. For example, the diagnostic concordance of bipolar disorder is 40-50%2), while the lifetime morbidity of bipolar disorder is around 1%. However, we have not fully understood the biological background of this high heritability. Specific genetic information and biological mechanisms to the onset of psychi-atric disorders remain unknown.Genetic information is coded as sequences of bases in DNA. All the creatures on Earth adopt this system. Homo sapiens is not an exception. After the first proposal of the double helix struc-ture of DNA as media of genetic information, researchers have been long pursuing what specific features of DNA contribute to the phenotypes of the individual. In psychiatry, the phenotypes in the individuals correspond to the diagnosis of psychi-atric disorders and their symptoms. Psychiatric researchers desire to know specific characters of DNA that contribute to psychiatric disorders. The initial effort of this endeavor started from single nucleotide polymorphisms (SNPs). SNP is one type of genetic variant as a conversion from one base to another with a frequency of one percent or more in the general population. This was a good start point because SNP is relatively easy to detect. However, the initial effort to hunt disease-related SNPs was not fruitful3). This effort is like fishing in the Pacific Ocean, and fishing all over the Pacific Ocean was not realistic at that time. To explore all over the Pacific Ocean, we needed a different approach.The birth of genomic analysisThe Human Genome Project was started in 1990 by the government of the United States. This project aimed at the complete catalog of the human genome4), which spans around three Giga base pairs as one haplotype. Empowered by the Human Genome Project and subsequent development of large-scale DNA sequencing technology such as SNP chip and next-generation sequencing5), comprehensive human genome analysis became realistic. This new approach is called “genomics.” The word “genome” means a whole set of genes, consisting of “gene” and “-ome” (a suffix meaning totality). In contrast to fishing in one spot, this approach can be compared to broad image capture by artificial space satellites. Capturing the images all over the Pacific Ocean became possible. Researchers are now armed with genomic technol-ogies to investigate the human genome, finding a lot of variants associated with medical diseases. As the cost of genomic technologies declines, the sample size of genomic research increases, and the conclusion from the analysis becomes robust. Psychiatric genomics is not an exception.A variant in the human genome receives different natural selection pressure with the phenotypes associated with it. Variants associated with severe diseases such as life-threatening pediatric cardio-vascular diseases tend to be negatively selected, thus having a lower frequency in the general popu-lation. In contrast, variants scarcely associated with severe diseases are neutral and can have a high frequency in the general population by genetic drift. To put it the other way around, variants with higher frequency in the general population (common variants) tend to have minor effects on diseases; variants with lower frequency in the general popu-lation (rare variants) tend to have a more signifi-cant effect on diseases6). Figure 1 illustrates the theoretical distribution of disease-associated vari-ants related to their effect on the disease and their frequency in the general population. Note that this illustration is a theoretical framework, and there are some exceptions to this framework, such as APOE and Alzheimer’s disease7, 8). In general, the contribution of common variants is supposed to be more significant to common diseases (e.g., diabetes mellitus) than rare diseases (e.g., Mendelian diseases); the contribution of rare variants is supposed to be more significant to rare diseases than common diseases. Disease-associated common variants have been detected mainly by genome-wide association studies (GWAS) using SNP chips. SNP chips can genotype millions of SNPs simultaneously, while the investigation is limited to the pre-designed 3