Future paths for integer programming and links to artificial intelligence
Computers and Operations Research - Special issue: Applications of integer programming
Computers and Operations Research
Identifying gene regulatory networks from experimental data
Parallel Computing - new trends in high performance computing
Tabu Search
DS '02 Proceedings of the 5th International Conference on Discovery Science
A mathematical program to refine gene regulatory networks
Discrete Applied Mathematics
A GRASP metaheuristic for microarray data analysis
Computers and Operations Research
Hi-index | 0.04 |
Gene regulatory networks are a common tool to describe the chemical interactions between genes in a living cell. This paper considers the Weighted Gene Regulatory Network (WGRN) problem, which consists in identifying a reduced set of interesting candidate regulatory elements which can explain the expression of all other genes. We provide an integer programming formulation based on a graph model and derive from it a branch-and-bound algorithm which exploits the Lagrangian relaxation of suitable constraints. This allows to determine lower bounds tighter than CPLEX on most benchmark instances, with the exception of the sparser ones. In order to determine feasible solutions for the problem, which appears to be a hard task for general-purpose solvers, we also develop and compare two metaheuristic approaches, namely a Tabu Search and a Variable Neighborhood Search algorithm. The experiments performed on both of them suggest that diversification is a key feature to solve the problem.