Computational Elucidation of Metabolic Pathways

Imran Shah

http://shah-lab.uchsc.edu

School of Medicine

University of Colorado

Elucidating the metabolic network of a living system is an important requirement for  modeling its physiological behaviour and for engineering its pathways. With the availability of whole genomes it is theoretically possible to infer the presence of putative enzymes and transporters in an organism. However, piecing this information into a complete picture is still mostly a daunting manual task for at least two reasons. First, we do not have   accurate and sufficient annotation of enzymatic function from sequence. Consequently, many proteins in completely sequenced microbes remain functionally uncharacterized. Second, inferring the causal biochemical connections within a metabolic network is not straightforward. We are developing a computational infrastructure to address these challenges. In earlier work we have developed a machine learning (ML) approach to improve the assignment of enzymatic function from sequence. More recently, we have developed an artificial intelligence (AI) approach for the prediction of metabolic pathways and their interactive visualization. In this talk I will present an overview of this work and its relevance to metabolic engineering.

Return to Agenda