henhao
Sign in
Systems BiologyNature Biotechnology
Peer-reviewed journal article

Memote: A community-driven effort towards a standardized genome-scale metabolic model test suite

Christian LieveniD, Moritz E. BeberiD, Brett G. OlivieriD, Frank BergmanniD, Meriç AtamaniD, Parizad BabaeiiD, Jennifer A. BartelliD, Lars M. BlankiD, Siddharth ChauhaniD, Kevin CorreiaiD, Christian DieneriD, Andreas DrägeriD, Birgitta E. EbertiD, Janaka N. EdirisingheiD, José P. FariaiD, Adam M. FeistiD, Georgios FengosiD, Ronan M. T. FlemingiD, Beatriz García-JiméneziD, Vassily Hatzimanikatis AiD, Wout van HelvoirtiD, Christopher S. HenryiD, Henning HermjakobiD, Markus J. HerrgårdiD, Ali KaafaraniiD, , Cheng ZhangiD

Abstract

Several studies have shown that neither the formal representation nor the functional requirements of genome-scale metabolic models (GEMs) are precisely defined. Without a consistent standard, comparability, reproducibility, and interoperability of models across groups and software tools cannot be guaranteed.\n\nHere, we present memote (https://github.com/opencobra/memote) an open-source software containing a community-maintained, standardized set of metabolic model tests. The tests cover a range of aspects from annotations to conceptual integrity and can be extended to include experimental datasets for automatic model validation. In addition to testing a model once, memote can be configured to do so automatically, i.e., while building a GEM. A comprehensive report displays the models performance parameters, which supports informed model development and facilitates error detection.\n\nMemote provides a measure for model quality that is consistent across reconstruction platforms and analysis software and simplifies collaboration within the community by establishing workflows for publicly hosted and version controlled models.

PDFDOICode
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
509
FWCI
22.86
Field percentile
top 1%
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 3 (11 Jul 2018) · CC BY 4.0

Memote, open-source Python software, runs a community-maintained, partly scored test suite (consistency, biomass, annotations, optional experimental data) on SBML-encoded genome-scale metabolic models, once or through continuous integration.

Genome-scale metabolic models (GEMs) are shared in inconsistent formats, and earlier surveys cited here found energy-generating cycles and biomass errors that distort predicted growth, plus missing annotations that hamper model exchange. Cycle checks exist (COBRA Toolbox functions, a modified Global Fit algorithm) but are not applied consistently; memote offers one community-maintained test suite that runs on SBML Level 3 FBC models and can retest every edit through continuous integration.

Key results

  1. ClaimedMemote supports two workflows. A one-off run produces a human-readable snapshot report that editors or reviewers can use. A reconstruction workflow pairs version control with continuous integration, so every tracked edit to a model is retested.

    One-off runs work through a web service or the command line. In the reconstruction workflow, results are saved for each model version and assembled into a commit-history report, which appears on the project's GitHub pages when continuous integration is enabled. A functional diff report compares two models side by side. No example run is shown. · Abstract; Memote: Community-driven quality control; Figure legends

  2. ClaimedCore tests are split into two sections. The scored section is designed to be independent of organism, model complexity and identifier type, so any two models can be compared by score. The unscored section holds statistics and context-specific checks such as biomass quality control.

    Four areas: basic (core components, metabolite formulas and charges, gene-protein-reaction rules, metabolic coverage), biomass (precursor production across conditions, consistency, non-zero growth), consistency (stoichiometric inconsistency, spurious energy metabolite production, blocked reactions) and annotation (MIRIAM cross-references, one primary ID namespace, Systems Biology Ontology terms). · Description of the test library; Figure legends

  3. ClaimedUsers can add experimental tests based on gene perturbation data, especially gene essentiality. To set the experimental condition (medium, flux balance analysis objective, known regulatory effects), they can optionally write YAML configuration files instead of custom scripts.

    Accepted inputs are CSV, TSV, XLS or XLSX. The authors argue that configuration files avoid the software rot of unmaintained scripts because they depend only on memote, and that keeping the model, configurations and data together in a versioned repository makes condition-specific models reproducible. No validation run against experimental data is reported. · Description of the test library; Figure legends

  4. ClaimedThe authors propose SBML Level 3 with the Flux Balance Constraints (FBC) package as the shared, tool-neutral format for GEMs, and make it memote's required input.

    SBML Level 3 Core covers reactions, metabolites and annotations, but not every component of a parameterised FBA model. FBC adds tool-neutral, validatable encoding of flux bounds, several linear objectives, gene-protein-reaction associations, metabolite formulas, charge and annotation. FBC Version 2 is called the de facto standard; implementation in software and adoption by repositories are deemed critical. · Introduction; SBML: Tool-independent model formulation

  5. ClaimedFor shared curation, memote writes the model as sorted YAML after each change so that GitHub, GitLab or BioModels can display line-by-line differences. A branching strategy lets model authors use the test results to gate external contributions.

    Functional differences between versions appear in a comparative diff report, and repositories can also be tested offline. A Python API and a web service are provided for integration. The authors plan links to the Department of Energy's Knowledge Base, the BiGG Database, BioModels and the RAVEN toolbox, plus an OptFlux plug-in with version 3.4. · Memote: Community-driven quality control; Discussion; Figure legends

Methods, data and reproducibility

Memote applies test-driven development, distributed version control and continuous integration to GEMs encoded in SBML Level 3 FBC. It runs a library of core tests (scored and unscored) plus optional tests against user-supplied experimental data, and renders HTML snapshot, diff and commit-history reports. The text describes the design only; no models are evaluated.

Methods & tools
memote, SBML Level 3 Flux Balance Constraints (FBC) Package, MetaNetX, identifiers.org, Systems Biology Ontology, MIRIAM annotations, YAML configuration files, GitHub, GitLab, BioModels, OptFlux (planned plug-in)
Code
https://github.com/opencobra/memotehttps://memote.readthedocs.io/en/latest/https://memote.dd-decaf.eu/https://github.com/opencobra/memote/wiki
Software & environment
Open-source Python package under the Apache License, Version 2.0. Input must be SBML Level 3 FBC. It is available as a command-line tool, a Python API and a web service. No package versions, containers or requirement files are stated.
Reproducibility gaps
No example models, reports or test results are provided. Outputs depend on external resources (MetaNetX, identifiers.org) that may change over time.
Limitations (6)
  • Memote's code is unit tested, but not every edge case can be foreseen, and it relies on external resources (MetaNetX, identifiers.org) that are likely to change. False-positive and false-negative results are therefore possible, and the authors advise reading reports critically. · stated by the authors
  • The tests cover stoichiometric models only. Models of metabolism and expression (ME-models), kinetic models and systems pharmacology models are out of scope, though the authors think the principles may transfer. · stated by the authors
  • Differences in modeling approach and organism make GEMs hard to compare, for example condition-constrained networks versus unconstrained metabolic databases. The scored section is described only as an attempt to make such models comparable. · stated by the authors
  • Version control of a single SBML file is only one possible design; database-backed tools such as Pathway Tools and AuReMe take a different approach. Tests based on multi-omics data are not yet supported. · stated by the authors
  • No GEM is actually run through memote in the paper. It presents no scores, example reports or runtimes, and no check that the tests catch the problems listed in the Introduction (e.g., energy-generating cycles). · noted in this summary
  • The text does not say how the scored tests are weighted or combined into a score, so the paper does not let a reader judge what a given score means. · noted in this summary

Who should read:Read this if you build, curate, review or distribute genome-scale metabolic models and want automated, version-controlled quality checks on SBML files, or are choosing a model format and test set to standardize on.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISAnton '23Gilbert '19Scharm '20Matveishina '25Sarathy '20Bäuerle '23
74% semantic matchSystems Biology
Mihail Anton, Eivind Almaas, Famke Bäuerle+33 more·bioRxiv·Mar 21, 2023· 21 citations