Approved research

Elaboration and validation of prediction models for chronic disease risk based on genomic, health and medical record data

Lay summary

As the increase in the incidence and mortality in chronic diseases is a serious problem all over the world, it is very important to predict and prevent chronic diseases in terms of public health. Disease onsets are generally caused by an interaction of genetic and environmental factors. Recent advanced techniques in lifelogging, genetic analysis and big data analytics make it easier to collect individual generated health-related data, and integration of these data with clinical features allow us to predict disease risk more accurately. For these reasons, this project aims to elaborate our genetic risk prediction models for chronic disease by considering health and medical record data using machine learning technique and to validate the effects of intervention based on these models with high-risk population cohorts (i.e. potential metabolic syndrome patients and cancer survivors). We will try to handle various diseases, however, clinical validation may be possible for only a few diseases. In terms of public health, a predicted disease risk based on many types of modifiable and unmodifiable factors can be used in healthcare services which encourage people to change their health behavior. Especially, it could motivate people in high-risk population to participate actively in health promotion. We ultimately expect that disease risk prediction models based on UK Biobank data will contribute to monitoring individuals' health status and delaying onset of diseases eventually.