Indian Journal of Science and Technology

Abstract

The DNA microarray technologies allow researchers to characterize the expression profiles of genes for related samples. The understanding of microarray data has led to the development of new methods in the detection of differentially expressed genes. Therefore, a powerful approach needs to be developed to deal with this interest to ensure it is a good detection. This technique is at first employed in individual gene analysis but recently it has been applied to gene set analysis. The relationship between genes in a gene set is analysed using Hotelling’s T² as a multivariate test statistic. However, the disadvantage of this test is the number of samples is larger than the number of variables when used in microarray studies. Consequently, the sample covariance matrix is not positive definite and singular, thus it cannot be inverted. This study explores the potential of the shrinkage approach to estimate the covariance matrix in Hotelling’s T² particularly when the high dimensionality problem occurred. The Hotelling’s T² statistic is combined with a shrinkage approach as an alternative estimation to estimate the covariance matrix, which detects significant gene sets. The performances of the proposed methods were assessed using simulation study. Shrinkage covariance matrix approach shows a promising result for detection of differentially expressed gene sets as compared to other methods.
Keywords: Differential Gene Expression, Gene Set Analysis, Hotelling’sT2, Shrinkage Covariance Matrix