A particle-based approach for topology estimation of gene networks

In this paper, an iterative particle-based method is proposed for topology estimation of gene networks. Using a particle filter for each gene expression, the connections among genes and the gene expressions are modeled by random measures. The probabilities of the possible topologies are computed using only estimates of gene expressions which allow for proposals of new topologies in an iterative manner. The resampling step of particle filtering eliminates the topologies with smaller weights and improves the results. The algorithm is compared with the Least Absolute Shrinkage and Selection Operator. The simulation results of the proposed method show better performance in capturing the interactions among genes.