Multi-Objective Approach for Protein Structure Prediction
This work proposes to optimize Protein Structure Prediction (PSP) using multi-objective ab initio approach. This paper addresses an application of modified NSGA-II (MNSGA-II) by incorporating controlled elitism and Dynamic Crowding Distance (DCD) strategies in NSGA-II for PSP by minimizing free Potential Energy (PE) and minimizing Solvent Accessible Surface area (SAS). In this model, a trigonometric representation is used to compute backbone and side-chain torsion angles of protein atoms. Free energy is calculated using Chemistry at HARvard Macromolecular Mechanics (CHARMm -22). SAS is calculated using dssp program. Both objectives together evaluate the structures of protein conformations. The evolution of protein conformations is directed by optimization of protein energy and surface area contributions using MNSGA-II. To validate the Pareto-front obtained using MNSGA-II, reference Pareto-front is generated using multiple runs of single objective optimization (RGA) with weighted sum of objectives. TOPSIS technique is applied on obtained non-dominated solutions to determine Best Compromise Solution (BCS). Result of MNSGA-II is compared with NSGA-II. The proposed model is validated with Met-enkephalin, a benchmark protein, obtaining very promising results.
KeywordsProtein Structure Prediction Free Potential Energy Solvent Accessible Surface Area modified NSGA-II Pareto-front Best Compromise Solution
Unable to display preview. Download preview PDF.
- 3.Unger, R., Moult, J.: Genetic algorithms for protein folding simulations. Biochim. Biophys. 231, 75–81 (1993)Google Scholar
- 5.Coello, C.A., Van Veldhuizen, D.A., Lamont, G.B.: Evolutionary Algorithms for Solving Multi-Objective Problems. Kluwer Academic Publishers (2002)Google Scholar
- 7.Calvo, J.C., Ortega, J.: Parallel protein structure prediction by multiobjective optimization. Parallel, Distributed and Network-based Processing 12(4), 407–413 (2009)Google Scholar
- 8.Shi, S.Y.M., Suganthan, N.: Parallel protein structure prediction by multiobjective optimization. KanGAL Report 2004007, 1–7 (2004)Google Scholar
- 10.Chamorro, A.E.M., Divina, F., Aguilar-Ruiz, J.S., Cortés, G.A.: A multi-objective genetic algorithm for the Protein Structure Prediction. In: Intelligent Systems Design and Applications ISDA 2011, pp. 1086–1090 (2011)Google Scholar
- 11.NIH Center for Macromolecular Modeling, & Bioinformatics, http://www.ks.uiuc.edu/
- 13.Centre for Molecular and Biomolecular Informatics, http://swift.cmbi.ru.nl/gv/dssp
- 15.Luo, B., Zheng, J., Xie, J., Wu, J.: Dynamic crowding distance – a new diversity maintenance strategy for MOEAs. In: Proceedings of the IEEE International Conference on Natural Computation, pp. 580–585 (2008)Google Scholar
- 18.Deb, K., Tewari, R., Dixit, M., Dutta, J.: Finding trade-off solutions close to KKT points using evolutionary multi-objective optimization. IIT Kanpur, KanGAL Report Number 2007006 (2007)Google Scholar
- 19.Biological Macromolecular Resource, Protein Data Bank, http://www.pdb.org/
- 21.An open-source molecular visualization system, http://www.pymol.org/