Title :
Feature selection for SNP data based on Relief-SVM
Author :
Zhang, Wenbin ; Wu, Yue ; Lei, Zhou ; Liu, Zongtian
Author_Institution :
Dept. of Comput. Eng. & Sci., Shanghai Univ., Shanghai, China
Abstract :
The presence of Single nucleotide polymorphism causes DNA sequence difference, affects protein changing the structure and function, which causes the human genetic disease. The whole genome wide association is a new strategy to screen SNP with disease related, it avoids the hypothesis before non-fully evidence, but it increases the difficulties of screening SNP with disease. Now in the whole genome wide association the association analysis of single SNPs has been considered and without the consideration of the interaction among SNPs, thus the SNP of screening are not entirely credible. In order to improve the classification accuracy of SNP selecting, this paper proposed a feature selection method based on Relief-SVM for SNP data, which can screen important SNP with disease related.
Keywords :
DNA; biology computing; diseases; genetics; genomics; molecular biophysics; polymorphism; proteins; support vector machines; DNA sequence; SNP data; feature selection; genome wide association; human genetic disease; protein; relief-SVM; single nucleotide polymorphism; Accuracy; Bioinformatics; Diseases; Genomics; Kernel; Silicon; Support vector machines; SNP; feature selection; genome wide association;
Conference_Titel :
Biomedical Engineering and Informatics (BMEI), 2011 4th International Conference on
Conference_Location :
Shanghai
Print_ISBN :
978-1-4244-9351-7
DOI :
10.1109/BMEI.2011.6098608