Inferring gene regulatory networks from expression data with prior knowledge by linear programming

Inferring gene regulatory networks from gene expression data is an important task in biological studies. In this work, we proposed an optimization model to infer regulatory relations among the functional genes from expression data based on the structural sparsity and/or prior knowledge. Specifically, we achieved the structural sparsity of the network by implementing a linear programming model, which also satisfies the conditions of the existing knowledge. The gene regulatory network is reconstructed by enforcing the sparse linkages with the consistency to the prior knowledge. The effectiveness of the method are demonstrated by several simulated experiments.