QSAR MODELING OF ANTIMICROBIAL PEPTIDES BASED ON THEIR STRUCTURAL AND PHYSICOCHEMICAL PROPERTIES
YONGZHONG LU *
Department of Biology, Qingdao University of Science and Technology, 266042, Qingdao, China
QIAN QIU
Department of Biology, Qingdao University of Science and Technology, 266042, Qingdao, China
DAOLE KANG
Department of Biology, Qingdao University of Science and Technology, 266042, Qingdao, China
JIE LIU
Department of Biology, Qingdao University of Science and Technology, 266042, Qingdao, China
*Author to whom correspondence should be addressed.
Abstract
Antimicrobial peptides (AMPs) have been considered as potential therapeutic agents against drug-resistant bacterial pathogens. Accurately relating the quantitative properties of AMPs with the activity may expedite novel AMP design. In this study, the structural and physicochemical features of peptides computed by the updated PROFEAT web server were applied to develop QSAR models for 101cationic peptides (CAMEL-s). The variables were optimized by stepwise multiple regression (SMR) and genetic algorithm (GA), QSAR models were constructed by multiple linear regression (MLR) and partial least squares (PLS), respectively. These models show good performance in reliability and predictability. The results indicate that these peptide descriptors may be promising in QSAR research of AMPs, and the QSAR models should be useful for novel AMP design.
Keywords: Antimicrobial peptides, peptide descriptor, amino-acid descriptor, quantitative structural activity relationship, genetic algorithm, multiple linear regression, partial least squares