・Developing novel technique for genome analysis using machine learning
In the 2000s, many genome-wide association analysis (GWAS) has been conducted, and associations between common variants, such as single nucleotide polymorphisms (SNPs) in the human genome, and diseases and traits have been uncovered. However, rare genetic variants, for example, which are found in only one in 10,000 people, are difficult to detect with conventional GWAS.
We are developing and examining a new analysis method using machine learning for rare variants detected by the next-generation sequencer. It is expected to clarify the relationships between rare variants and various diseases, which have not been detected by the conventional statistical method.
・Developing Novel Health Indicator Using Artificial Intelligence
A variety of tests are performed in medical practice, including physical examinations, blood tests, and medical imaging studies. However, in actual practice, it is sometimes only a piece of that information that is used for diagnosing disease or deciding on a treatment plan. For example, respiratory physicians tend to focus on the lung-related aspects, while cardiologists tend to focus only on test results related to cardiac function.
We are using artificial intelligence to create new health indicators that can be used to estimate the presence of disease, the degree of aging, and the risk of developing disease in the future from the combination of various test results and genomic information. We aim to create actionable indicators that will lead to early detection of diseases and early therapeutic interventions.