Search bioRxiv⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/2023.02.08.527721

A Structured Evaluation of Genome-Scale Constraint-Based Modeling Tools for Microbial Consortia

Scott, W. T.·Benito Vaquerizo, S.·Zimmermann, J.·Bajic, D.·Heinken, A.·Suarez-Diez, M.·Schaap, P. J.
View original source ↗Export metadata ↓

Abstract

Harnessing the power of microbial consortia is integral to a diverse range of sectors, from healthcare to biotechnology to environmental remediation. To fully realize this potential, it is critical to understand the mechanisms behind the interactions that structure microbial consortia and determine their functions. Constraint-based reconstruction and analysis (COBRA) approaches, employing genome-scale metabolic models (GEMs), have emerged as the state-of-the-art tool to simulate the behavior of microbial communities from their constituent genomes. In the last decade, many tools have been developed that use COBRA approaches to simulate multi-species consortia, under either steady-state, dynamic, or spatiotemporally varying scenarios. Yet, these tools have not been systematically evaluated regarding their software quality, most suitable application, and predictive power. Hence, it is uncertain which tools users should apply to their system and what are the most urgent directions that developers should take in the future to improve existing capacities. This study conducted a systematic evaluation of COBRA-based tools for microbial communities using datasets from two-member communities as test cases. First, we performed a qualitative assessment in which we evaluated 24 published tools based on a list of FAIR (Findability, Accessibility, Interoperability, and Reusability) features essential for software quality. Next, we quantitatively tested the predictions in a subset of 14 of these tools against experimental data from three different case studies: a) syngas fermentation by C. autoethanogenum and C. kluyveri for the static tools, b) glucose/xylose fermentation with engineered E. coli and S. cerevisiae for the dynamic tools, and c) a Petri dish of E. coli and S. enterica for tools incorporating spatiotemporal variation. Our results show varying performance levels of the best qualitatively assessed tools when examining the different categories of tools. The differences in the mathematical formulation of the approaches and their relation to the results were also discussed. Ultimately, we provide recommendations for refining future GEM microbial modeling tools. Author summaryConstraint-based modeling employing genome-scale reconstructions of microbial species has become one of the most successful approaches for studying, analyzing, and engineering microbial consortia. Over the past decade, many constraint-based modeling tools have been published to examine an immense variety of microbial consortia spanning from the application areas of bioremediation to food and health biotechnology. However, new potential users lack an overview of the quality and performance of existing metabolic modeling tools that would guide their choice. To tackle this issue, we examined 24 tools for genome-scale metabolic modeling of microbial consortia. After an initial qualitative screening, we quantitatively evaluated 14 adequate tools against published experimental data that included different organisms and conditions. We conducted simulations and evaluated model features such as predictive accuracy, computational time, and tractability in capturing critical physiological properties. We found that, generally, more up-to-date, accessible, and documented tools were superior in many important aspects of model quality and performance. Although, in some cases, we observed tradeoffs in older, less elaborate tools that can be more accurate or flexible. This work has broad implications to help researchers navigate the most suitable tools, and suggests to developers opportunities for improvement of the currently existing capabilities for metabolic modeling of multi-species microbial consortia.

Source connections

Explore related subjects

systems biologyPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Scott, W. T., Benito Vaquerizo, S., Zimmermann, J., Bajic, D., Heinken, A., Suarez-Diez, M., Schaap, P. J.. 2023-02-09. A Structured Evaluation of Genome-Scale Constraint-Based Modeling Tools for Microbial Consortia. https://doi.org/10.1101/2023.02.08.527721

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Low-fidelity simulator selection for multi-fidelity inference of epithelial vertex models

Vertex models link cell division, neighbour exchange, and intercellular forces to epithelial tissue form through parameters such as junctional line tension and perimeter contractility. Bayesian inference of these parameters from data quantifies their uncertainty, but for stochastic models of proliferating tissues it requires many costly simulations. Multi-fidelity simulation-based inference reduces this cost by training on many runs of a cheaper, less accurate ('low-fidelity') simulator before refining with a few full ('high-fidelity') runs. Low-fidelity versions can be constructed in several ways, for example by stopping simulations early, coarsening the timestep, or averaging fewer replicates, but it is unclear which to choose before committing to costly training. To address this, we introduce PRISM (Paired Ranking Index for Simulator Matching), a screen that compares each candidate with the full simulator on paired runs, scoring agreement in cell-packing statistics against cost. For a vertex model of a proliferating tissue, PRISM flags stopping early as unreliable before training; once trained, this choice can produce confidently wrong parameter estimates. By contrast, a coarser timestep, within the range tested, is cheap and accurate; while fewer replicates are accurate but too costly. Using a screened low-fidelity simulator, multi-fidelity inference outperforms approximate Bayesian computation and neural-network inference trained on full simulations alone, at matched cost and with well-calibrated uncertainty. The same pattern recurs for a second vertex model fitted to published Drosophila pupal-wing imaging data, although there stopping halfway proves safe. Together, our findings suggest that which approximations are safe depends on the model, and that screening them before training makes multi-fidelity inference of tissue mechanics practical.

systems biology↗
bioRxiv2026

Calibrating developmental dynamics from single-cell atlases: lineage paths outrun least-action routes

Single-cell atlases infer developmental dynamics from destructive snapshots, yet how far realized lineage paths depart from such inferred dynamics has never been measured with controls. We introduce LADD (Least-Action Developmental Dynamics), a measurement framework that learns a stochastic developmental vector field from time-series single-cell RNA-seq and computes minimum-action and minimum-time paths, together with a three-layer calibration protocol that separates genuine path deviation from the generic cost of centroid measurement, and we compare its forecasting component with optimal-transport and flow-matching methods. In the lineage-resolved Caenorhabditis elegans atlas, the raw action ratio of observed lineage-centroid trajectories to optimal paths is 3.79, yet forward-simulated ensembles of the learned model itself score 2.07, leaving a genuine excess of about two-fold. In zebrafish (Danio rerio), the observed ratio sits entirely at the model's floor. Three mechanisms that could explain the residual deviation (detours, flow-opposition, and fate-map complexity) are each excluded; what remains is speed, with lineages moving 30-43% faster along the learned flow. We report validity boundaries for temporal resolution, cross-species comparison, and energetic interpretation, delimiting where action-based trajectory comparisons are meaningful.

systems biology↗
bioRxiv2026

Choice of annotation data significantly impacts automated reconstruction of genome-scale metabolic networks

Genome-scale metabolic models are valuable tools for studying metabolism in many applications, such as metabolic engineering or understanding human disease. However, curated models are available for only a limited number of organisms, and generating high-quality models for additional organisms remains a challenge. To help address this challenge, several software tools have been developed that can produce draft models in an automated fashion, but the quality of these draft models lags behind manually curated ones. Thus, there is still a need to improve these tools. Previous work has focused primarily on developing better methods for each step in the reconstruction process, such as gene annotation and gap-filling. Here, we focus on the data that underlies these methods, such as known connections between protein sequence and function. We find that changing the source of reconstruction data has a substantial impact on the contents and predictions of generated models. In fact, the source of data can have a bigger impact than the choice of reconstruction method. Additionally, we show that the reconstruction process is sensitive to the accuracy of gene annotations in the reconstruction data. Finally, we demonstrate the importance of having functional annotation data from an organism that is closely related to the one for which a model is being built. These observations highlight the importance of reconstruction data and identify principles that should be considered when selecting data for reconstruction tools.

systems biology↗