Inferring Gene Regulatory Networks with Sparse Bayesian Learning and phi-mixing coefficient

A Gene Regulatory Network (GRN) is a graphical representation of how genes within a cell regulate each other via various mechanisms. Inferring a GRN from high-throughput experimental data is an important problem in systems biology. In this paper, a new algorithm is presented for inferring a GRN from steady-state gene expression data. The new algorithm utilizes a Sparse Bayesian Learning (SBL) approach and a measure of dependence between random variables called the φ-mixing coefficient. To evaluate the performance of the algorithm, it is compared with two state of the art algorithms on several synthetic datasets. The results demonstrate that our algorithm compares favorably with these two algorithms on moderate sized networks.