452036 Prediction of Bacterial Fluxomics Using Machine Learning and Constraint Programming

Monday, November 14, 2016
Grand Ballroom B (Hilton San Francisco Union Square)
Gang Wu, Enery, Environmental and Chemical Engineering Department, Washington University in St. Louis, Saint Louis, MO, Tolutola Oyetunde, Energy, Environmental and Chemical Engineering, Washington University in St. Louis, St. Louis, MO, Wu Jiang, Boxed Inc., Kazuyuki Shimizu, Institute of Advanced Biosciences, Keio University,, Yinjie Tang, Enery, Environmental and Chemical Engineering, Washington University in St. Louis, St Louis, MO and Forrest Sheng Bao, University of Akron

Metabolic information is important for disease treatment, bioprocess optimization, environmental remediation, biogeochemical cycle regulation, and our understanding of life’s origin and evolution. 13C-MFA can quantify microbial physiology at the level of metabolic reaction rates. Mining the relationship between environmental and genetic factors and metabolic fluxes hidden in existing fluxomic data will lead to predictive models that can significantly accelerate flux quantification. In this paper, we present a web-based platform (MFlux: http://mflux.org/) that predicts the bacterial central metabolism via machine learning, leveraging data from ~100 13C-MFA papers on heterotrophic bacterial metabolisms. Three machine learning methods, namely Support Vector Machine (SVM), k-Nearest Neighbors, and Decision Tree, were employed to study the sophisticated relationship between influential factors and metabolic fluxes. We performed a grid search of the best parameter set for each tested algorithm and verified their performance through 10-fold cross validation. SVM yielded the highest accuracy of all three algorithms. Further, we employed quadratic programming to adjust flux profiles to satisfy stoichiometric constraints. Multiple tests have shown that MFlux can predict fluxomes as a function of bacterial species, substrate types, growth rates, oxygen conditions, and cultivation methods. 

In summary, aided by constraint programming and quadratic optimization, our machine learning platform can predict meaningful metabolic information about bacterial species in their environments. Further, it can offer constraints to improve the accuracy of flux balance analysis. This study infers that the bacterial metabolic network has a certain degree of rigidity in allocating carbon fluxes, and different microbial species may share common regulatory strategies for balancing carbon and energy metabolisms. As a proof of concept, we demonstrate that the use of data driven models (e.g., artificial intelligence) may assist mechanistic based models to elucidate the topology of microbial fluxomes.


Extended Abstract: File Not Uploaded
See more of this Session: Poster Session: Bioengineering
See more of this Group/Topical: Food, Pharmaceutical & Bioengineering Division