Implementation, Validation and Profiling of a Genetic Algorithm for Molecular Conformational Optimization

Gueorguiev, Victor and Kuttel, Michelle (2016) Implementation, Validation and Profiling of a Genetic Algorithm for Molecular Conformational Optimization, Proceedings of Proceedings of the Annual Conference of the South African Institute of Computer Scientists and Information Technologists (SAICSIT 2016), September 2016, Johannesburg, ACM New York, NY, USA.

PDF a16-gueorguiev.pdf Download (2MB)

Abstract

Prediction of the lowest energy conformation of a protein chain is a challenging optimization problem in computational chemistry and biology. Simple lattice-based protein models have been shown to be effective representations of the characteristics of proteins important in protein folding. An effective genetic algorithm for conformational optimization of proteins represented by the hydrophobic-hydrophillic lattice model was recently published. In this work, we create a publically available implementation of this genetic optimization algorithm. Tests of our implementation show equivalent performance to that reported for the original, in terms of both optimal conformation and number of function evaluations. In addition, we test our implementation across a range of data set sizes to characterize the performance of the algorithm as chain length increases: benchmarking that is necessary for future optimization and parallelization of the algorithm.

Item Type:	Conference paper
Uncontrolled Keywords:	Genetic algorithm; HP Lattice; conformational search; energy optimization; hydrophobic-hydrophilic model
Subjects:	Applied computing > Physical sciences and engineering Computing methodologies > Modeling and simulation
Date Deposited:	24 Oct 2016
Last Modified:	10 Oct 2019 15:32
URI:	http://pubs.cs.uct.ac.za/id/eprint/1102

Actions (login required)

View Item