AbstractsBiology & Animal Science

As the metabolic network reconstruction process via manual curation proves to be a tedious process, the need for an automated reconstruction tool becomes imminent. There has been previous attempts to automate metabolic network reconstruction. These existing approaches rely on sequence similarity to infer functions. Consequently, they fail to identify functions in the absence of sequence similarity or when sequences are heterogeneous. We propose a new method to automate metabolic network reconstruction that combines sequence similarity information and the likelihood of a certain enzyme to catalyze an enzymatic transformation. At the end, we combine both sequence homology and functional homology information to automatically reconstruct metabolic network.