A minimum classification error framework suitable for multicriteria gene selection: discovery of differentially methylated genes in small B-cell lymphomas. (3rd February 2010)

• Record Type: Journal Article
• Authors: Popescu, Mihail; Arthur, Gerald; Rahmatpanah, Farahnaz; Sjahputera, Ozy; Shi, Huidong; Caldwell, Charles; Keller, James
• Is Part Of: International journal of computational intelligence in bioinformatics and systems biology. Volume 1:Number 3 (2010)
• Page Start: 274
• Page End: 296
• Publication Date: 2010-02-03
• Subjects: gene methylation -- minimum classification errors -- K-nearest neighbour -- multicriteria gene selection -- B-cell lymphomas -- DNA -- differentially methylated genes; Bioinformatics -- Periodicals; Computational intelligence -- Periodicals; Systems biology -- Periodicals; 572.8028563
• Languages: English
• ISSNs: 1755-8034

Abstract

Gene methylation appears to influence cellular activity by silencing the expression of certain genes believed to be responsible for significant biological processes, including neoplastic transformation. In this paper, we present a framework for the identification of differentially methylated genes in several types of small B-cell lymphomas (SBCLs) using DNA differential methylation microarrays. The proposed approach uses the patient classification error to tune various steps of the algorithm such as the data normalisation, the gene ranking and the threshold for gene selection. We present the results obtained by application of the proposed framework on two gene methylation datasets: one with 43 samples and three SBCL types and another with 38 samples and two SBCL subtypes. Some of the identified genes are known to be involved in critical pathways such as apoptosis and proliferation while others function as tumour suppressor genes or oncogenes.