Classification method for prediction of multifactorial disease development using interaction between genetic and environmental factors

Multifactorial disease such as life style related diseases, for example, cancer, diabetes mellitus, myocardial infarction (Ml) and others, is thought to he caused by complex interactions between polygenic basis and various environmental factors. In this study, we used 22 polymorphisms on 16 candidate genes that have been characterized and potentially associated with MI in terms of biological function and 6 environmental factors. To predict development for MI and classify the subjects into personally optimum development patterns, we extracted risk factor candidates (RFCs) composed of state which is a derivative form of polymorphisms and environmental factors using statistical test and selected risk factors from RFCs using Criterion of Detecting Personal Group (CDPG) defined in this study. We could predict development of blinded data simulated as unknown their development more than 80% accuracy and identify their causal factors using CDPG.