Reconstructing genetic networks with concurrent representation of instantaneous and time-delayed interactions

Although living organisms can have some genetic interactions occurring instantaneously while others with time-delay, current modeling techniques for genetic network reconstruction make simplifications and assume that the interactions can be either of these but not both. In this paper, we propose a gene regulatory network reconstruction algorithm that can model concurrent occurrence of both, instantaneous as well as time-delayed interactions, thus providing a better representation of the original biological processes. First we introduce a novel framework using the Bayesian network (BN) formalism that can model both types of interactions. A gene regulatory network reconstruction algorithm using this proposed framework is then developed that employs an evolutionary search strategy and a decomposable scoring metric based on information theoretic quantities. Investigations of our approach are performed using both, the synthetic data as well as Saccharomyces cerevisiae gene expression data. Comparisons with recent reconstruction methods show the superiority of the proposed method.