Title: A group-specific tuning parameter for hybrid of SVM and SCAD in identification of informative genes and pathways

Authors: Muhammad Faiz Misman; Mohd. Saberi Mohamad; Safaai Deris; Siti Zaiton Mohd. Hashim

Addresses: Artificial Intelligence & Bioinformatics Research Group, Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia, Johor, Malaysia ' Artificial Intelligence & Bioinformatics Research Group, Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia, Johor, Malaysia ' Artificial Intelligence & Bioinformatics Research Group, Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia, Johor, Malaysia ' Soft Computing Research Group, Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia, 81310, Skudai, Johor Darul Takzim, Malaysia

Abstract: The pathway-based microarray classification approach leads to a new era of genomic research. However, this approach is limited by the issues in quality of pathway data. Usually the pathway data are curated from biological literatures and in specific biological experiment (e.g., lung cancer experiment), context free pathway information collection process takes place leading to the presence of uninformative genes in the pathways. Many methods in this approach neglect these limitations by treating all genes in a pathway as significant. In this paper, we proposed a hybrid of support vector machine and smoothly clipped absolute deviation with group-specific tuning parameters (gSVM-SCAD) to select informative genes within pathways before the pathway evaluation process. Our experiment on canine, gender and lung cancer datasets shows that gSVM-SCAD obtains significant results in identifying significant genes and pathways and in classification accuracy.

Keywords: data mining; smoothly clipped absolute deviation; SCAD; support vector machines; SVM; gene selection; bioinformatics; group specific tuning parameters; gene identification; pathway identification; informative genes; significant genes; significant pathways; classification accuracy.

DOI: 10.1504/IJDMB.2014.064013

International Journal of Data Mining and Bioinformatics, 2014 Vol.10 No.2, pp.146 - 161

Received: 03 Oct 2011
Accepted: 02 May 2012
Published online: 21 Oct 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article