Causal Modeling of Gene Regulatory Network

The analysis of high-throughput experimental data, such as microarray gene expression data, is currently seen as a promising way of finding regulatory relationships between genes. Network inference algorithms are powerful computational tools for identifying putative causal interactions among variables from observed data. In this paper, we propose a network reconstruction technique to predict not only the structure but also the direction and sign of regulation using a genetic algorithm (GA). The networks consisting of nodes (genes), directed edges (gene-gene interactions) and dynamics of regulation are assigned scores using the presented causal model based on partial correlation. The highest scoring network best fits the expression data. As GAs are stochastic, the algorithm is repeated several times and the final network is reconstructed by combining the most significant connections identified from the high scoring networks. The presented technique is applied to the well known Saccharomyces cerevisiae microarray dataset and the reconstructed network is observed to be consistent with the results found in literature