A genetic algorithm encoded with the structural information of amino acids and dipeptides for efficient conformational searches of oligopeptides. Issue 13 (2nd February 2016)

• Record Type: Journal Article
• Authors: Ru, Xiao; Song, Ce; Lin, Zijing
• Is Part Of: Journal of computational chemistry. Volume 37:Issue 13(2016)
• Page Start: 1214
• Page End: 1222
• Publication Date: 2016-02-02
• Subjects: dihedral angle -- structural prediction -- conformational coverage -- potential energy surface -- geometry optimization; Chemistry -- Data processing -- Periodicals; 542.85
• DOI: 10.1002/jcc.24311
• Languages: English
• ISSNs: 0192-8651

Abstract

Abstract : The genetic algorithm (GA) is an intelligent approach for finding minima in a highly dimensional parametric space. However, the success of GA searches for low energy conformations of biomolecules is rather limited so far. Herein an improved GA scheme is proposed for the conformational search of oligopeptides. A systematic analysis of the backbone dihedral angles of conformations of amino acids (AAs) and dipeptides is performed. The structural information is used to design a new encoding scheme to improve the efficiency of GA search. Local geometry optimizations based on the energy calculations by the density functional theory are employed to safeguard the quality and reliability of the GA structures. The GA scheme is applied to the conformational searches of Lys, Arg, Met‐Gly, Lys‐Gly, and Phe‐Gly‐Gly representative of AAs, dipeptides, and tripeptides with complicated side chains. Comparison with the best literature results shows that the new GA method is both highly efficient and reliable by providing the most complete set of the low energy conformations. Moreover, the computational cost of the GA method increases only moderately with the complexity of the molecule. The GA scheme is valuable for the study of the conformations and properties of oligopeptides. © 2016 Wiley Periodicals, Inc. Abstract : The structural information of amino acids and dipeptides is carefully analyzed. An improved GA algorithm with a new encoding strategy utilizing the structural information is proposed. Technical improvements are also made to minimize the possibility of premature convergence of the GA search. Applications to representative amino acids, dipeptides and tripeptide with complicated side chains confirm that the new GA scheme is both efficient and reliable for providing the most complete conformational coverage of the molecules.