DocumentCode
2551676
Title
Sensitivity Analysis of Boosting PSI-Blast with Case Study on Subcellular Localization
Author
Mai, F ; Mak, M.W. ; Hung, Y.S. ; Kung, S.Y.
Author_Institution
Dept. of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong
fYear
2007
fDate
27-29 Aug. 2007
Firstpage
39
Lastpage
44
Abstract
This paper studies the sensitivity of PSI-BLAST with respect to the ´h´ parameter. Observing that the standard PSI-BLAST is sensitive to parameter ´h´ in the high-value region, we propose a new technique, called boosting PSI- BLAST, to reduce the sensitivity. By constraining ´h´ to a small value first so as to reduce the chance of early corruption and then relaxing it gradually to increase divergence, the boosting PSI-BLAST not only can reduce the sensitivity to h-value, but also may strike a good balance between corruption and divergence in profiles. Tests on Reinhardt and Hubbard´s eukaryotic protein dataset verify that our method is better in reducing the sensitivity of profile alignment scores to h-value than the standard PSI-BLAST.
Keywords
biology computing; cellular biophysics; molecular biophysics; proteins; sensitivity analysis; PSI-BLAST database searching program; boosting PSI-BLAST; eukaryotic protein dataset; h-value; profile alignment scores; protein classification; sensitivity analysis; subcellular localization; Accuracy; Boosting; Computer aided software engineering; Databases; Process control; Proteins; Sensitivity analysis; Signal processing; Testing;
fLanguage
English
Publisher
ieee
Conference_Titel
Machine Learning for Signal Processing, 2007 IEEE Workshop on
Conference_Location
Thessaloniki
ISSN
1551-2541
Print_ISBN
978-1-4244-1566-3
Electronic_ISBN
1551-2541
Type
conf
DOI
10.1109/MLSP.2007.4414279
Filename
4414279
Link To Document