Guiding the Refinement of Biochemical Knowledgebases with Ensembles of Metabolic Networks and Machine Learning

Abstract

Mechanistic models are becoming common in biology and medicine. These models are often more generalizable than datadriven models because they explicitly represent biological knowledge, enabling simulation of scenarios that were not used to construct the model. While this generalizability has advantages, it also creates a dilemma: how should model curation efforts be focused to improve model performance? Here, we develop a machine learningguided solution to this problem for genomescale metabolic models. We generate an ensemble of candidate models consistent with experimental data, then perform in silico ensemble simulations for which improved predictiveness is desired. We apply unsupervised and supervised learning to the simulation output to identify structural variation in ensemble members that maximally influences variance in simulation outcomes across the ensemble. The resulting structural variants are high priority candidates for curation through targeted experimentation. We demonstrate this approach, called A utomated M etabolic M odel E nsemble D riven E limination of U ncertainty with S tatistical learning ( AMMEDEUS ), by applying it to 29 bacterial species to identify curation targets that improve gene essentiality predictions. We then compile these curation targets from all 29 species to prioritize refinement of the entire biochemical database used to generate them. AMMEDEUS is a fully automated, scalable, and performancedriven recommendation system that complements human intuition during the curation of hypothesisdriven models and biochemical databases. SignificanceMechanistic computational models, such as metabolic and signaling networks, are becoming common in biology. These models contain a comprehensive representation of components and interactions for a given system, making them generalizable and often more predictive than simpler models. However, their size and connectivity make it difficult to identify which parts of a model need to be changed to improve performance further. Here, we develop a strategy to guide this process and apply it to metabolic models for a set of bacterial species. We use this strategy to identify model components that should be investigated, and demonstrate that it can improve predictive performance. This approach systematically aides the curation of metabolic models, and the databases used to construct them, without relying on the intuition of the curator.

Community standards to facilitate development and address challenges in metabolic modeling

“…Tests evaluating network topology can be used to evaluate the more subjective features of the model, by using connectedness as a proxy for inferring the scope of manual curation or the quality of a reconstruction. However, this requires such topological measures (and/or machine learning; Medlock & Papin, ) to be combined with biological knowledge of the system. Condition‐specific tests (often referred to as “metabolic tasks” in the COBRA field) are developed to evaluate the biological meaning of the network and attempt to represent specific biochemical experiments.…”

Transcriptomics data integration for context-specific modeling of Atlantic salmon metabolism: functional evaluation of methods based on metabolic tasks

“…from the same omics data, and predictions could be based on agreement among models in the ensemble (Biggs and Papin, 2017). Ensemble modeling could also help improve the models themselves by applying machine learning to their contents and predictions (Medlock and Papin, 2020). Indeed, recent studies have demonstrated context-specific ensemble modeling with a single MEM (Rodríguez-Mier et al ., 2021) and combined multiple MEMs to build a single model (Vieira et al ., 2022).…”

Synergisms of machine learning and constraint‐based modeling of metabolism for analysis and optimization of fermentation parameters

“…[ 154 ] ML methods such as ensemble methods have shown improvements in the refinements of genome‐scale metabolic reconstructions. [ 155 ] Third, although very accurate, the results of the integrated models are not necessarily appropriate for large‐scale industrial fermentations. This arises from the difference between fermentation conditions in lab‐scale and industrial‐scale bioreactors.…”