On the explicit use of enzyme-substrate reactions in metabolic pathway analysis
Document Type
Conference Proceeding
Date of Original Version
1-1-2018
Abstract
Flux balance (or constraint-based) analysis has been the mainstay for understanding metabolic networks for many years. However, recently Lucia and DiMaggio [1] have argued that metabolic networks are more correctly modeled using game theory, specifically Nash Equilibrium, because it (1) captures the natural competition between enzymes, (2) includes rigorous chemical reaction equilibrium thermodynamics, (3) incorporates element mass balance constraints, and therefore charge balancing, in a natural way, and (4) allows regulatory constraints to be included as additional constraints. The novel aspects of this work center on the explicit inclusion of enzyme-substrate reactions at the cellular length scale and molecular length scale protein docking information in metabolic network modeling. This multi-scale information offers the advantages of directly (1) computing cellular enzyme concentrations and activities, (2) incorporating genetic modification of enzymes, and (3) encoding the effects of age-related changes in enzymatic behavior (e.g., protein misfolding) within any pathway. Molecular length scale binding histograms are computed using protein-ligand docking and directly up-scaled to the cellular level. A small, proof-of-concept example from the Krebs cycle is presented to illustrate key ideas. Numerical results show that the proposed approach provides a wealth of quantitative enzyme information.
Publication Title, e.g., Journal
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume
10710 LNCS
Citation/Publisher Attribution
Lucia, Angelo, Edward Thomas, and Peter A. DiMaggio. "On the explicit use of enzyme-substrate reactions in metabolic pathway analysis." Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 10710 LNCS, (2018): 88-99. doi: 10.1007/978-3-319-72926-8_8.