A Preliminary Study on Identifying Probable Biomarker of Type 2 Diabetes using Recursive Feature Extraction

Abstract

Microarray technology has the ability to measure the level expression of thousand genes by single experiment and it can be used by biologist to study about the effect of treatments, disease and developmental stages on their expressions. Microarray based on gene expression profiling can be used to observing the response of expression genes to pathogens and identify which expressions genes are changed by comparing the expression in infected to that uninfected cells or tissue. Type 2 diabetes mellitus is a metabolic disorder that causes an increase in blood sugar due to decreased insulin secretion by pancreatic beta cells and insulin disorder (insulin resistance). The number of incidences of diabetes mellitus in Indonesia reached 10 million and 53% from the patients do not realized that they are infected and 90% case of diabetes from whole world is type 2 of diabetes. Therefore, in this paper, we identify probable biomarker of type 2 diabetes using microarray based on gene expression data. But the risk of using microarray data is the large dimension of data so have to find a way how to solve that problem to get a good prediction result. In this paper will use recursive feature extraction for predicting biomarkers of diabetes mellitus type 2 from microarray gene expression data.