Source-linked AI summary
One file to share them all: Using the COMBINE Archive and the OMEX format to share all information about a modeling project
Frank T. Bergmann, Richard Adams, Stuart Moodie, Jonathan Cooper, Mihai Glont, Martin Golebiewski, Michael Hucka, Camille Laibe, Andrew K. Miller, David P. Nickerson, Brett G. Olivier, Nicolas Rodriguez, Herbert M. Sauro, Martin Scharm, Stian Soiland-Reyes, Dagmar Waltemath, Florent Yvon, Nicolas Le Novère
TL;DR
Computational modeling experiments require many interdependent descriptions, but sharing and maintaining those files can hinder reproducibility. The paper presents OMEX and the COMBINE Archive, a ZIP-based single-file package with manifests, metadata, and model-related files. The archive supports exchanging complete modeling projects, with tools available and substantial compression demonstrated for a large reconstruction.
Problem
Computational experiments depend on multiple model, simulation, data, and procedural descriptions, while managing separate files can impede reliable reuse and reproduction.
Method
The paper defines OMEX, a ZIP-based format for COMBINE Archives containing a manifest, optional metadata, and files encoded in standard or Internet Media Types.
Results
23MB versus 230MB demonstrates a 90% compression ratio for an archive containing the expanded Recon2.1 human metabolic reconstruction.
Takeaways & Limitations
A single file containing protocols and data supports easier model reuse, improved reproducibility, and tracking of model development.
Takeaways & Limitations
OMEX does not yet provide a well-defined general mechanism for linking elements across different files, such as SBGN-ML glyphs to SBML elements.
Abstract
from arXiv · showhide
Background: With the ever increasing use of computational models in the biosciences, the need to share models and reproduce the results of published studies efficiently and easily is becoming more important. To this end, various standards have been proposed that can be used to describe models, simulations, data or other essential information in a consistent fashion. These constitute various separate components required to reproduce a given published scientific result. Results: We describe the Open Modeling EXchange format (OMEX). Together with the use of other standard formats from the Computational Modeling in Biology Network (COMBINE), OMEX is the basis of the COMBINE Archive, a single file that supports the exchange of all the information necessary for a modeling and simulation experiment in biology. An OMEX file is a ZIP container that includes a manifest file, listing the content of the archive, an optional metadata file adding information about the archive and its content, and the files describing the model. The content of a COMBINE Archive consists of files encoded in COMBINE standards whenever possible, but may include additional files defined by an Internet Media Type. Several tools that support the COMBINE Archive are available, either as independent libraries or embedded in modeling software. Conclusions: The COMBINE Archive facilitates the reproduction of modeling and simulation experiments in biology by embedding all the relevant information in one file. Having all the information stored and exchanged at once also helps in building activity logs and audit trails. We anticipate that the COMBINE Archive will become a significant help for modellers, as the domain moves to larger, more complex experiments such as multi-scale models of organs, digital organisms, and bioengineering.
4 Current affiliation: Eight Pillars Ltd, 19 Redford Walk, Edinburgh EH13 0AG
The listed affiliations include the University of Oxford, the University of Washington, the University of Auckland, and the VU University Amsterdam.
- The University of Oxford affiliation is associated with the Department of Computer Science.
- The University of Auckland affiliation is associated with the Auckland Bioengineering Institute.
- The VU University Amsterdam affiliation is associated with Systems Bioinformatics.
11 Department of Bioengineering, University of Washington, Seattle, WA, 98195
The paper motivates a unified archive for increasingly complex computational modeling projects, whose models, simulations, datasets, and procedures are distributed across multiple files and standards.
- Motivation: Reproducing computational experiments requires descriptions of methods, tools, substrates, setups, environmental influences, and post-processing steps.
- Existing standards: Structured formats such as SBML, CellML, and NeuroML support exchange of model structure, while SBRML and SED-ML describe results and simulation setups.
- Problem: Increasingly complex systems use modular models assembled from multiple files, with datasets and simulation procedures integral to model descriptions.
- Problem: Managing multiple files is tedious and error-prone because files can move or disappear, while model changes must be propagated across reused instances.
- Contribution: The COMBINE initiative developed an archive enabling exchange of all information required to reproduce a modeling project.
Format of the archive
The COMBINE Archive uses OMEX to package modeling-project files in a ZIP container that can combine multiple standard formats and other information.
- Archive encoding: OMEX encodes the COMBINE Archive as a ZIP file using the .omex extension by default.
- File extensions: Archives may use extensions such as .sedx, .sbex, .cmex, .sbox, .neux, and .phex to indicate a principal standard format.
- File extensions: The indicated extension is only a convenience because one archive may contain files in several standard formats, including both SBML and CellML.
- Archive contents: A COMBINE Archive may contain descriptions of models, simulations, graphical representations, metadata, and other information.
Manifest file
The manifest identifies every archive file and its format and location, while optional metadata supplies provenance and administrative information for reproducibility and audit trails.
- Manifest file: Every COMBINE Archive requires a root-level manifest.xml containing a flat list of all archive files, their types, and their locations.
- Manifest entries: Manifest entries specify relative locations and formats using COMBINE identifiers, Internet Media Types, or other registered representations.
- Manifest entries: The optional master attribute marks the main content to be read or presented first, such as a SED-ML simulation or a hierarchical master model.
- Metadata: An RDF-based metadata file can describe the archive, creator, creation date, last update, software, and external references.
- Metadata: Metadata about archive contents helps build action logs and audit trails, including for models used in drug development.
RESULTS AND DISCUSSION
The COMBINE Archive specification was finalized as OMEX version 1, and implementation support became available in libraries and modeling software.
- The archive specification emerged from extending SED-ML’s original focus on simulation descriptions and required models to encompass any useful file type.
- OMEX version 1 was released on September 15, 2014, after two release candidates and a comment period.
- Several independent libraries and modeling tools, including PySCeS, VCell, and Tellurium, support the COMBINE Archive.
- Table 1 catalogs implementations supporting the COMBINE Archive.
Use cases
The COMBINE Archive supports diverse modeling workflows by packaging models, simulations, datasets, and results for reproducible exchange across applications and computational nodes.
- Combining multiple information types in one file supports researchers across varied modeling and simulation applications.
- Systems biology models: The archive packages all documents describing models and associated simulations, including multiple models or modular model files.
- Drug discovery models: Drug-discovery workflows require coordinated exchange of datasets, model code, outputs, plots, diagnostics, and accompanying descriptions.
- Large hybrid modular models: Large hybrid models may combine different modeling approaches and simulation tools whose results influence one another.
- Automatic machine-only transfer: Archives can transfer simulation jobs and results between web servers, back-end nodes, and applications without human interaction.
Related efforts
The COMBINE Archive builds on earlier single-package and research-object approaches while choosing a simpler, more flexible structure than several industry archive standards.
- JSim Project Files pioneered packaging notes, model code, and control parameters for a simulation experiment.
- The Research Object Bundle focused on preserving scientific experiments in data-intensive science.
- OMEX shares ZIP-based archival concepts with UCF, Open Document Format, and EPUB Open Container Format.
- The initial OMEX specification prioritizes simplicity and flexibility over the relative complexity and rigidity of industry standards.
General discussion
The archive improves packaging, transfer, and storage of modeling projects, but OMEX still lacks a general mechanism for relationships between elements in different files.
- OMEX does not currently provide a standard mechanism for describing relationships between elements across different files.
- No well-defined method currently links SBGN-ML glyphs with corresponding SBML elements, although mapping solutions have been discussed.
- A Recon2.1 archive shrinks approximately 230 MB of contained files to 23 MB, a 90% compression ratio.
- Keeping all components inside the archive avoids uncertainties about external links and file-version consistency, though external access could be more flexible.
- Distributing protocols and data in one file is expected to ease model reuse, improve result reproducibility, and support tracking model development.
AVAILABILITY AND REQUIREMENTS
The COMBINE Archive specification is available online, and supporting data are included with the article and its additional file. The relevant standards include formats for models, simulations, results, annotations, and graphical notation.
- The COMBINE Archive specification is available through the identifiers.org OMEX specification link.
- The article’s supporting data are included within the article and its additional file.
- The listed standards include COMBINE, MIASE, MIRIAM, NuML, OCF, ODF, OMEX, OOXML, and OPC.
- The standards list also includes SBGN, SBGN-ML, SBML, SBRML, SED-ML, UCF, and URI.
AUTHORS’ CONTRIBUTIONS
The initial COMBINE Archive version was written by FTB, RA, and NLN, while other contributors helped develop the specification and implementations. Two example archives provide packaged modeling resources.
- AUTHORS’ CONTRIBUTIONS: FTB, RA, and NLN wrote the initial version of the COMBINE Archive.
- AUTHORS’ CONTRIBUTIONS: JC, MG, MH, AKM, SM, DPN, SSR, NR, HMS, and DW contributed to the current specification.
- AUTHORS’ CONTRIBUTIONS: SM, MG, BO, MS, FY, and SSR developed implementations, and all authors contributed to the manuscript.
- EXAMPLE ARCHIVES: repressilator.omex contains a Repressilator model encoded in SBML and a simulation experiment description as supporting information.
- EXAMPLE ARCHIVES: recon2-1-x.omex contains an expanded version of the Recon1.1 model.