Reconstruction of gene regulatory network based on gravitation field algorithm

Ming ZHENG,Gui-xia LIU,You ZHOU,Chun-guang ZHOU
DOI: https://doi.org/10.13229/j.cnki.jdxbgxb201402024
2014-01-01
Abstract:In order to resolve the low accuracy and inefficiency of reconstruction of Gene Regulatory Networks (GRNs) in system biology, we proposed a novel inference algorithm from gene expression data based on differential equation model. In this algorithm, two methods are employed for inferring GRNs. One is Singular Value Decomposition (SVD) method and the other one is Gravitation Field Algorithm (GFA). The SVD method is used to decompose gene expression data, determine the algorithm solution space, and get all candidate solutions of GRNs. The GFA is the kernel part of the proposed algorithm. The GFA is divided into four parts: initialization, division of solution space, movement operator and absorption. Random group method is used in division of solution space. Every element movement method is used in movement operator. The proposed algorithm is validated on both the simulated scale-free network and real benchmark gene regulatory network in network database. Both genetic algorithm and simulated annealing are also used to evaluate GFA. The cross-validation results confirm the effectiveness of the proposed algorithm, which outperforms significantly other existing algorithms.
What problem does this paper attempt to address?