Heuristic-Based Ant Colony Optimization Algorithm For Protein Functional Module Detection In Protein Interaction Network
Loading...
Date
2017-07
Authors
Sallim, Jamaludin
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
Ant colony optimization (ACO) is a metaheuristic algorithm that has been successfully applied to several types of optimization problems such as scheduling, routing, and more recently for solving protein functional module detection (PFMD) problem in protein-protein interaction (PPI) networks. For a small PPI data size, ACO has been successfully applied to but it is not suitable for large and noisy PPI data, which has caused to premature convergence and stagnation in the searching process. To cope with the aforementioned limitations, we propose two new enhancements of ACO to solve PFMD problem. First, we combine ACO with nearest neighbor heuristic (termed ACOPFMD-NN) that utilized the candidate lists as a selection strategy used by artificial ants when they construct the solution. Second, we apply the information theory concept, information entropy combined with ACO (termed ACOPFMD-IE) to handle the path selection by controlling two important parameters of the ACO; pheromone trail and heuristic information . The experiments on a gold standard benchmark dataset “Saccharomyces cerevisiae” from two popular databases DIP and MIPS has shown that our two enhancements have improved the performance of basic ACO, two recent metaheuristics and state-of-the-art of PFMD algorithms. In terms of quantitative results, ACOPFMD-NN has improved the accuracy up to 67% (DIP), 80% (MIPS) while ACOPFMD-IE has improved the accuracy up to 73.8% (DIP), 87.3% (MIPS). In terms of qualitative result, ACOPFMD-NN has improved the accuracy up to 32% (DIP), 33% (MIPS) while ACOPFMD-IE has improved the accuracy up to 69% (DIP), 59% (MIPS). ACOPFMD-IE has also obtained a better accuracy over two metaheuristics algorithms; 80% (DIP – compared with IGA algorithm), 74% (MIPS – compared with ABC-IFC algorithm).
Description
Keywords
Heuristic-based ant colony optimization algorithm , functional module detection in protein interaction network