BMB Rep. 2013; 46(1): 041-046  
Partial AUC maximization for essential gene prediction using genetic algorithms
Kyu-Baek Hwang1, Beom-Yong Ha1, Sanghun Ju1 & Sangsoo Kim2,*

1School of Computer Science and Engineering, Soongsil University, 2School of Systems Biomedical Science, Soongsil University, Seoul 156-743, Korea

Correspondence to: Tel: +82-2-820-0457; FAX: +82-2-820-0816, E-mail: sskimb@ssu.ac.kr
Received: August 2, 2012; Revised: August 8, 2012; Accepted: August 8, 2012; Published online: January 31, 2013.
© Korean Society for Biochemistry and Molecular Biology. All rights reserved.

Abstract
Identifying genes indispensable for an organism's life and their characteristics is one of the central questions in current biological research, and hence it would be helpful to develop computational approaches towards the prediction of essential genes. The performance of a predictor is usually measured by the area under the receiver operating characteristic curve (AUC). We propose a novel method by implementing genetic algorithms to maximize the partial AUC that is restricted to a specific interval of lower false positive rate (FPR), the region relevant to follow-up experimental validation. Our predictor uses various features based on sequence information, protein-protein interaction network topology, and gene expression profiles. A feature selection wrapper was developed to alleviate the over-fitting problem and to weigh each feature's relevance to prediction. We evaluated our method using the proteome of budding yeast. Our implementation of genetic algorithms maximizing the partial AUC below 0.05 or 0.10 of FPR outperformed other popular classification methods.
Keywords: AUC, Classification, Essential genes, Genetic algorithms, Partial AUC


This Article

e-submission

Archives