An Algorithm for the Discovery of Phenotype Related Metabolic Pathways

Microorganisms are being increasingly used in industrial processes due to certain beneficial phenotypes they exhibit. Improving the ability of microorganisms to exhibit these phenotypes has driven interest in identifying the genes that are responsible for a given phenotype. Some of these phenotypes are the result of various chemical compounds being modified by a series of metabolic reactions, or metabolic pathways, catalyzed by specific enzymes. Recently, comprehensive, generic metabolic networks have been defined, which describe possible ways in which certain chemical compounds may be modified by known metabolic reactions. In this paper, we aim to discover phenotype related metabolic pathways by identifying subnetworks of a generic metabolic network that are highly conserved in phenotype expressing organisms and rarely conserved in non-phenotype expressing organisms. To do this, we introduce a graph search algorithm that finds and expands highly conserved seed networks based on their evolutionary bias towards phenotype expressing organisms. We hypothesize that the evolutionarily conservation of these subnetworks in phenotype expressing organisms is likely due to the fact that they represent metabolic pathways responsible for the expression of the phenotype. We test our approach using aerobic and anaerobic organisms to identify pathways related to aerobic respiration. We find that the pathways identified by our algorithm are found primarily in aerobic organisms and that metabolic pathways known to be related to aerobic respiration are covered by the pathways identified by our algorithm. We finish by discussing the ongoing and future work related to this methodology.