The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title:	Ranking single nucleotide polymorphisms with support vector regression in continuous phenotypes
Author:	Shahidain, Seif
View Online:	njit-etd2011-082 (xii, 44 pages ~ 3.5 MB pdf)
Department:	Department of Mathematical Sciences
Degree:	Master of Science
Program:	Computational Biology
Document Type:	Thesis
Advisory Committee:	Roshan, Usman W. (Committee chair) Wei, Zhi (Committee member) Dhar, Sunil Kumar (Committee member)
Date:	2011-05
Keywords:	Support vector machines Single nucleotide polymorphisms Continuous phenotypes
Availability:	Unrestricted
Abstract:	Support vector machines (SVM) have been used to improve the ranking of single nucleotide polymorphisms (SNPs) over traditional chi-square tests in disease case studies [2]. In this investigation, ranking SNPs with support vector regression (SVR) was compared to the Wald test in predicting continuous phenotypes. SVR-ranked SNPs consistently outperformed the Wald test-ranked SNPs to provide a more accurate prediction of the phenotype with fewer SNPs across several methods of prediction.