Estimating local false discovery rates to identify the differentially expressed genes in microarrays

To detect differentially expressed genes (DEGs) in microarray experiments thousands of genes are tested against a null hypothesis. The false discovery rate (FDR) is a statistical concept for quantifying uncertainty during multiple testing. In this paper, a Hidden Markov Model (HMM) based approach is proposed to identify DEGs using local false discovery rates (Lfdr). We estimate the p-values and the proportion of the true null hypothesis using HMM, calculating the Lfdr. We assess the proposed method using four FDR-controlling procedure based methods. In terms of multiple hypothesis testing power, we demonstrate that our proposed method is more suitable to identify DEGs in microarrays. We also show the validity of the proposed method by applying it to a real microarray data set. © 2012 Binary Information Press.