(234g) Computational Pathway Identification and Strain Optimization for Biofuel Production

We present an integrated computational base to support pathway identification and strain optimization with an emphasis on biofuel production. An efficient graph-based algorithm is presented for the exhaustive identification of all pathways enabling the production of a targeted biofuel molecule. The algorithm searches over a database of biotransformations that spans reactions from KEGG, Metacyc, BRENDA and other resources with an emphasis on C4+ alcohols. The identified pathways are then integrated into the genome-scale model of the production host (e.g., Escherichia coli). We next describe the application of the OptFlux computational framework to pinpoint engineering modifications (knock-outs/up/down) that are required for the targeted biofuel overproduction. This is accomplished by classifying reactions (and combinations thereof) in the metabolic model depending upon whether their flux values must increase, decrease or become equal to zero to meet the pre-specified overproduction target. A ?force set? can then be extracted that contains a sufficient and non-redundant set of reactions that need to be directly changed to meet the production requirements. We apply the integrated framework for the production of 1-butanol, isobutanol, and other alcohols in E. coli using the most recent in silico E. coli model, iAF1260. The proposed computational workflow not only recapitulates existing pathways and engineering strategies but also reveals novel and non-intuitive ones that boost production by using and performing coordinated changes on sometimes distant pathways.