Peptide Drugs QSAR Modeling Based on a New Descriptor of Amino Acids- SVGT

Title:Peptide Drugs QSAR Modeling Based on a New Descriptor of Amino Acids- SVGT

Volume: 13 Issue: 3

Author(s): Jianbo Tong, Lingxiao Li, Shuling Liu and Jia Chang

Affiliation:

Keywords: Peptide drugs, amino acids, descriptor, QSAR, multiple linear regression, stepwise multiple regression.

Abstract: Quantitative structure-activity relationships (QSAR) approach is one of the recommended approaches in finding the relationship between molecular structures of amino acids and the activity of peptide drugs. In this work, variable screening by stepwise multiple regression (SMR) technique, a multiple linear regression (MLR) model was built with a new descriptor of amino acids-SVGT. Predictive ability and robustness of the models have been analyzed strictly by both training set and test set, with squared cross-validation correlation coefficient (Q_LOO²) and squared correlation coefficient between predicted and observed activities (r²). Moreover, the rm 2 (overall) metrics is used to further improve the predictive ability of the QSAR models. The obtained models with the Q_LOO 2, r² and r_m² (overall) were 0.943, 0.952 and 0.943 for antimicrobial peptides; 0.826, 0.978 and 0.713 for oxytocin; and 0.804, 0.956 and 0.786 for angiotensin- converting enzyme. Satisfactory results showed that SVGT can offer good account of relationships between activity and structure of peptide drugs.

Cite this article as:

Tong Jianbo, Li Lingxiao, Liu Shuling and Chang Jia, Peptide Drugs QSAR Modeling Based on a New Descriptor of Amino Acids- SVGT, Letters in Drug Design & Discovery 2016; 13 (3) . https://dx.doi.org/10.2174/1570180812999150819094745