Association of single nucleotide polymorphism and phenotype in type 2 of diabetes mellitus using Support Vector Regression and Genetic Algorithm

Abstract

Precision Medicine is used to improve proper health care and patients' quality of life, one of which is diabetes. Diabetes Mellitus (DM) is a multifactorial and heterogeneous group of disorders characterized by deficiency or failure to maintain normal glucose homeostasis. About 90% of all DM patients are Type 2 Diabetes Mellitus (T2DM). Biological characteristics and genetic information of T2DM disease were obtained by looking for associations in Single Nucleotide Polymorphism (SNP) which allows for determining the relationship between phenotypic and genotypic information and identifying genes associated with T2DM disease. This research focuses on the Support Vector Regression method and Genetic Algorithm to obtain SNPs that have previously calculated the correlation value using Spearman's rank correlation. Then do association mapping on the SNP results from the SVR-GA selection and check pastasis interaction. The results produced 14 SNP importance. Evaluation of the model using the mean absolute error (MAE) obtained is 0.02807. If the value of MAE is close to zero, then a model can be accepted. The genes generated from the association can be used to assist other researchers in finding the right treatment for T2DM patients according to their genetic profile.