With the sequencing and annotation of the entire genomes of numerous microorganisms has come the advent of the genome scale metabolic models such as the iJR904 model of E. coli from Palsson and coworkers. Metabolic flux anal. (MFA) is widely employed to generate the flux distributions that satisfy stoichiometric mass, energy, redox, and charge balance constraints in these models. However, not all of the distributions from MFA will be thermodynamically feasible. Some flux distributions will utilize pathways involving reactions that are thermodynamically unfavorable; these pathways will be incapable of producing desired metabolic end products or intermediates at an acceptable concn. This work focuses on assessing the thermodn. feasibility of the flux distributions generated by performing MFA on a genome scale metabolic model. The free energy change of every reaction in the model is estd. based on a group contribution method, and the alternative pathways to objectives such as growth or growth precursors are examd. for thermodn. bottlenecks. Flux distributions utilizing the most thermodynamically favorable reactions and pathways are discussed with respect to their physiol. significance. [on SciFinder (R)]