TY - GEN
T1 - Estimation of cross-hybridization signals using support vector regression
AU - Yijun, Sun
AU - Li, Liu
AU - Popp, Mick
AU - Farmerie, William
PY - 2006
Y1 - 2006
N2 - Microarray technology is a powerful biotechnology tool which allows researchers to simultaneously evaluate the expression of thousands of genes, if not the entire expressed genome, of an organism. Measures of gene expression are determined by the differential hybridization of labeled mRNA from experimental samples to DNA probes affixed to the array. The accuracy of these measurements is influenced by the binding specificity between the labeled samples and the probes. Evaluating the level of cross-hybridization is therefore critically important in obtaining accurate measures of gene expression. In this paper we present a support vector regression based predictor that utilizes the nucleotide content of the DNA probes as a means for estimating the level of cross-hybridization. Experimental results from three microarray data sets are presented. Our results indicate that we can identify genes when the measured fluorescent signal values are less than those predicted from cross-hybridization. In these cases we do not consider the genes to be expressed.
AB - Microarray technology is a powerful biotechnology tool which allows researchers to simultaneously evaluate the expression of thousands of genes, if not the entire expressed genome, of an organism. Measures of gene expression are determined by the differential hybridization of labeled mRNA from experimental samples to DNA probes affixed to the array. The accuracy of these measurements is influenced by the binding specificity between the labeled samples and the probes. Evaluating the level of cross-hybridization is therefore critically important in obtaining accurate measures of gene expression. In this paper we present a support vector regression based predictor that utilizes the nucleotide content of the DNA probes as a means for estimating the level of cross-hybridization. Experimental results from three microarray data sets are presented. Our results indicate that we can identify genes when the measured fluorescent signal values are less than those predicted from cross-hybridization. In these cases we do not consider the genes to be expressed.
UR - http://www.scopus.com/inward/record.url?scp=33845596284&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33845596284&partnerID=8YFLogxK
U2 - 10.1109/IMSCCS.2006.58
DO - 10.1109/IMSCCS.2006.58
M3 - Conference contribution
AN - SCOPUS:33845596284
SN - 0769525814
SN - 9780769525815
T3 - First International Multi- Symposiums on Computer and Computational Sciences, IMSCCS'06
SP - 17
EP - 21
BT - First International Multi- Symposiums on Computer and Computational Sciences, IMSCCS'06
T2 - First International Multi- Symposiums on Computer and Computational Sciences, IMSCCS'06
Y2 - 20 April 2006 through 24 April 2006
ER -