Frequently Asked Questions
- What is BioModels Database and how does it differ from other resources?
- Citations and contact information
- Questions dealing with model correctness, reuse and distribution
- How reliable are the models hosted in BioModels Database?
- What should I do if I find an error in a model?
- What are the conditions of use and distribution for unmodified models originating from BioModels Database?
- What are the conditions of use and distribution for modified models that originated from BioModels Database?
- Can I convert a model from BioModels Database into another SBML version?
- Questions dealing with the provided features of BioModels Database
- Questions dealing with model submission
- How to submit a model?
- Can I submit a model before it is described in a published paper?
- Will a model be publicly accessible immediately after its submission?
- Can reviewers access unpublished models?
- What are the supported model encoding formats?
- How to submit a revised version of a model?
- Why does my SBML model contain no species or reactions after a conversion from CellML?
- What do I do if I receive error messages when trying to submit a model?
- Questions dealing with the curation of the models
- What checks are performed to ensure model correctness?
- What is MIRIAM?
- How is MIRIAM compliance ensured?
- What is annotation, and what purpose does it serve in a model?
- Who annotates the models?
- How are annotations stored in SBML?
- When is a model completely annotated?
- What are the tools used by the curators of BioModels Database?
- What is the non-curated branch of BioModels Database?
- Questions dealing with the underlying software infrastructure
- Miscellaneous questions
What is Biomodels Database?
BioModels Database is a repository of mathematical models representing biological systems. It currently hosts a range of models describing processes like signalling; protein-drug interaction interactions; metabolic pathways; epidemic models and many more. The models that BioModels hosts are usually described in peer-reviewed scientific literature and in some cases, they are generated automatically from pathway resources (Path2Models). These models are manually curated and semantically enriched with cross-references to external data resources (such as publications, databases of compounds and pathways, ontologies, etc.). BiModels allows the scientific community to store, search and retrieve mathematical models of their interest. In addition to offering support for supporting models in different formats and enabling their conversion between different representational formats, the database offers programmatic access to developers via web services.
Users can browse and search the content of the repository, and download models in SBML format, as well as various other formats, such as XPP, VCML, SciLab, Octave, BioPAX, PNG, SVG, ... A human readable summary of each model is also available in PDF.
More information can be found in the associated publications.
How does BioModels Database differ from other databases of models?
Unlike other pathway databases like Reactome, BioModels contains quantitative information representing the quantities (amount or concentration of species) or kinetics in a model. A formal model can merge several biochemical reactions into one, or conversely, it can contain reactions without counterparts in the corresponding biological context. Having a quantitative representative model allows a user to carry out simulations where a model produces quantitative results commensurate with the available experimental knowledge.
In addition to this, BioModels has its own controlled annotation and vocabulary which allows users to search not only for particular models based on their internal components elements, and extensive additional annotation added to the model through manual curation. These annotations increase access and visibility, both for the model and relevant resources linked to the model on BioModels, thereby facilitating the understanding of the concepts upon which the model is founded.
How to cite BioModels Database?
Please, have a look at the citation information page.
How to cite a model present in BioModels Database?
The best way to cite a model present in BioModels Database is to state the reference publication associated with the model. You can also mention the model's identifier (of the form "BIOMD" or "MODEL" followed by 10 digits).
How to contact the team behind BioModels Database?
The easiest way to contact the team developing and maintaining the software infrastructure and the content of BioModels Database is to use the following email address: biomodels-net-support AT lists.sf.net.
You can also refer to the information provided on the contact us page.
I wish to contact the original author(s) of a model, but the listed email does not work. What should I do?
This will happen occasionally as people move on to new roles or to new positions in different institutions. There is no easy solution for this, so we would suggest in the first instance to write to the BioModels Database curation team, which can have them look into it on your behalf. Please provide as much information as you have available, and avenues you may have explored already. The curation team can be contacted via: biomodels-cura AT ebi.ac.uk.
How reliable are the models hosted in BioModels Database?
Before being publicly available on BioModels Database, a model passes through a stringent curation pipeline. This ensures its syntactic correctness, semantic soundness, and its correspondence with its reference publication, both in terms of model structure and simulation results. Consequently the structure of a model would not normally change, while its annotation is expected to improve constantly over time.
What should I do if I find an error in a model?
The models present in BioModels Database have already been extensively checked and corrected. However, it remains possible that some errors may have crept through our rigorous curation pipeline. If you discover any errors with a specific model, or have any potential concerns, please do contact us. Comments and bugs for specific models may be submitted directly from the menu bar at the top of the web page describing each model, where you will find a "Submit Model Comment/Bug" link.
What are the conditions of use and distribution for unmodified models originating from BioModels Database?
What are the conditions of use and distribution for modified models that originated from BioModels Database?
You can modify and freely distribute a modified version a model that is present in BioModels Database, in whole or part. A modified model is defined as one where an existing model is extended/reduced/merged with other models. In this case, there is a change in the biological description and components.
The modified model must be renamed, and a new submission to BioModels must be made. This new model will then undergo the curation process again as a separate model. When curating the modified model, a curator will also add a reference to the parent model on BioModels from which it was derived.
Can I convert a model from BioModels Database into another SBML version?
Conversion of SBML model files between different levels/versions is necessary for users who use tools that do not support the versions/levels of the original SBML file. The current version of BioModels does not provide an automatic conversion between different SBML levels/versions. You can do this offline using tools like COPASI.
In the next release of BioModels, we aim to provide automatic SBML convertors which will allow you to download the model file in different SBML versions.
How to browse and search BioModels Database?
One of the main classification of models on BioModels is based on their curation status i.e. manually curated and non-curated models. By clicking on Browse, you can choose curated or non-curated models, along with being able to apply several filters to identify your models of interest. This includes filtering on the basis of resources like CHEBI and Uniprot, disease associations, taxonomy and annotations like GO.
A link to Browse is available from the home page as well as the always visible top menu bar. Once you have identified a model of interest, you can click on the model identifier to reach the model description page.
What information can be found using the model tabs?
Here is a brief description of the information available through the tabs available on each curated model's description page:
- Overview: This view has the abstract of the paper describing the model, citation and contributor details. It also has model level annotations, curation and validation status of the model and certification comments from the curator.
- Files: This view has a list of all the available files associated with the model. This includes the most recent SBML model file, biopax file, SBGN representation of the model (.png file), a model report file (.pdf), .xpp file and a few other commonly used formats.
- History: This view provides information about the original submitter, submission date and details regarding different versions of the model, in case changes were made to the original submitted model.
- Curation: This view shows curation comments and the reproduced simulation figure from the paper, which was generated by the curator during the manual curation process.
How to download a model from BioModels Database?
There are several way to download the models:
- Each model can be downloaded from its own description page, via a "Download Model" link on the side bar menu which is active only when examining a model's description page and on all its tabs.
- All the models can be downloaded in one single archive. Several archives are available: two per release (one with only the SBML files, one with the SBML and all the export files) and one automatically generated every week (only contains the SBML files). The archives are available on the FTP server.
- SBML version of the models can be obtained using the Web Services.
Does BioModels Database provide SBGN maps?
BioModels Database provides graphical representations of the models in various formats (PNG and SVG). These files are available from the "Files" tab on a model's description page.
Due to workforce limitations, and the volume of work that generating SBGN exports necessitate, the layout of most of these maps is not currently fully compliant with the specification of SBGN Process Diagrams Level 1 Version 1. However, the BioModels.net team has been involved in SBGN development since it's inception, and is fully committed to support it as much, and as quickly, as possible.
Does BioModels Database export models under MATLAB format?
BioModels Database developers are doing their best to provide a variety of export formats. BioModels Database is a public resource, and is totally committed to support open standard formats. Therefore, our first priority is the support of free (non-proprietary) software. In addition, BioModels Database is entirely funded by taxpayers' money, and we feel it is inappropriate to develop software to interface with commercial tools as a priority.
That said, we do provide exports for Octave, which uses a m-file similar to the one used by MATLAB, so you should be able to use it in this environment. Additionally, there exist several packages offering SBML support for MATLAB. See the SBML Toolbox, the Systems Biology Toolbox, and simbiology, the toolbox developed by MathWorks.
How to submit a model?
Anybody can submit a model to BioModels Database by creating an account on BioModels. Even if you are not the actual author or encoder of a model you can submit it for inclusion in BioModels Database. If you do so, you will be recorded as the submitter, not as an author or an encoder. However, we do encourage you to contact the original authors to be sure they are happy with the submission.
If a model is encoded in SBML or CellML, the submission is entirely processed, via the submission page. If the model is encoded using a different format, please read the section of this FAQ about the supported formats. If the model can be converted into SBML or CellML, please contact us.
Once the model has been submitted, the submitter will receive an email notification with a unique, stable and perennial submission identifier. This identifier is composed of the string "MODEL" followed by 10 digits, for example MODEL1003250000 (these digits are actually based on the submission date and time). This identifier can be used to reference the model in publications, and once the model is publicly available on BioModels Database, users can directly access the model using this identifier.
BioModels Database currently only publishes models which have been described in a peer reviewed scientific publication. Therefore, authors are encouraged to submit models before publication of the associated paper (and will receive an identifier that they can used in the publication), but the models will only be publicly available on BioModels Database once the paper has been published.
In order to access a model, knowing it submission identifier, you can use the URL: http://www.ebi.ac.uk/biomodels/ followed by the identifier. For example: https://www.ebi.ac.uk/biomodels/MODEL1002160000.
Can I submit a model before it is described in a published paper?
Yes. Models can be submitted prior to the publication of the associated paper(s). It is actually strongly advised to do so: at the submission time, each model is assigned a unique and perennial identifier which allows users to access and retrieve it. This identifier can be used by authors as a reference in their publications.
If you submitted a model before the publication of the associated paper, please keep us informed of the status of your publication. That way, we can make the model available from BioModels Database as quickly as possible.
When submitting your paper, we encourage you to include the submission identifier in your publication and cite BioModels by referring to the guidelines on how to cite BioModels Database.
Will a model be publicly accessible immediately after its submission?
No. Models are not directly visible and retrievable by the public as soon as they are submitted to BioModels Database. From submission to their public release, all models undergo various automated and manually performed curation and annotation steps to ensure a consistent level of quality and accuracy.
Moreover, all models are only made publicly available from BioModels Database after the publication of the corresponding papers. Please refer to the section of this FAQ about the access to a model after its submission for more information about when a model will be publicly available from BioModels Database.
Finally, new models are usually made publicly available during "releases". There are between 2 to 4 releases of the database a year. Please contact us if you need your model to be available earlier.
Can reviewers access unpublished models?
BioModels Database can provide access to unpublished models to reviewers. This is not automatic, so the model submitter needs to request it (for example emailing us at biomodels-cura AT ebi.ac.uk). In this case, we will provide the submitter with a URL which can be used to download the model encoded in SBML. If security is a concern, we can supply a protected access, where reviewers will be asked for a password to access the model.
In future updates to BioModels, we plan to automate this process. If you have more specific needs, please do not hesitate to contact us.
What are the supported model encoding formats?
Currently BioModels Database only provides full support for a subset of the modelling formats landscape. This is mainly due to the fact that we strongly believe in interoperability, which requires distributing to the community models in semantically rich, standard formats. The full list of supported modelling formats is available on the submission page, and basically covers SBML (all Levels and Versions) , CellML (1.0 and 1.1), PharmML, Matlab, Mathematica and COMBINE Archive.
However, we are aware that those formats do not cover all modelling techniques used today. As we are committed to encourage models sharing and reuse, we will strive to find a solution to host models originally encoded in MATLAB or from other generic modelling platforms. So, if you cannot convert your model(s) to one of the fully supported formats, please do contact us (at biomodels-cura AT ebi.ac.uk, which is a non public mailing list dedicated to our curators).
How to submit a revised version of a model?
- If the model is not yet published in BioModels Database
You can send us the revised version at: biomodels-cura AT ebi.ac.uk together with the model submission identifier and a comment explaining which modifications were made. This mailing list is not public and the only recipients are the curators of BioModels Database, so you can safely send us models not yet published this way.
- The model has already been published in BioModels Databse
Any modification that would lead to a change in the biological component, results, etc. should be considered as new entry. We could accept revision to the model if the authors found an error in the publication, which they have reported/submitted the correction (errata) to the journal. For all other changes, the revised model is treated as a new entry, and curated before publishing.If you only made minor changes to the model, please follow the procedure described above (for not yet published models).
Why does my SBML model contain no species or reactions after a conversion from CellML?
The issue is that there is no annotation in the CellML file that could help the converter to determine if a certain variable is a SBML species, parameter or compartment. And the same applies to any CellML equation.
The current version of the converter is not trying to do any complicated guesses and therefore converts every CellML variable into a SBML parameter and all the equations into rules. We are currently trying to define some annotations that would allow to create better SBML files from CellML and vice-versa, the ultimate goal would be to be able to do a round-trip conversion, starting either from a SBML model or a CellML model, without loosing any information.
What do I do if I receive error messages when trying to submit a model?
Any validation issues detected during the submission process will be reported as warnings, making it possible to proceed with the submission in spite of the encoding problems. It may be perfectly possible to submit a model that is incorrectly encoded in SBML and then share it with someone that can help fix the SBML representation.
What checks are performed to ensure model correctness?
Before being made publicly available through BioModels Database, models are thoroughly verified. This involves several types of checks, some automated and some manually performed by a team of curators. The following checks are performed:
- The syntax of the encoded file is verified. This implies a set of correct hierarchy of elements and attributes, a consistent use of identifiers, but also a deeper verification of the consistency of the model depending on its format (e.g. a parameter controlled by a rule cannot be declared constant in SBML).
- The correspondence between the encoded model and the model described in the published article is checked. This implies verification of the structure of the reaction network, the formalisms used, the values of the parameters, the amounts and the units.
- Finally the model is simulated to see if the published results can be reproduced. For this step, a different simulator is employed than the one used by the original authors, in order to ensure the reproducibility of the results.
These checks come from the MIRIAM set of guidelines.
What is MIRIAM?
MIRIAM is the Minimum Information Required in the Annotation of Models (publication and press Release). Initiated by the BioModels.net project, it is a set of guidelines defining how a model should be encoded and annotated in order to be successfully distributed, exchanged and ultimately reused.
In particular MIRIAM requires that a model provides all the information necessary to instantiated in a simulation, such as the initial conditions. In addition, the reaction graph generated from this simulation must reproduce the results of the original publication, in which the model was first described. Moreover, the MIRIAM guidelines require that all model components contain sufficient controlled annotation such that each component can be unambiguously identified. The MIRIAM Registry and Identifiers.org have been developed specifically to provide generation and resolution services for unique and perennial identifiers to be used in controlled annotations.
All models stored in the curated branch of BioModels Database are MIRIAM-compliant.
How is MIRIAM compliance ensured?
All the models in the curated branch are fully MIRIAM compliant. Each of the MIRIAM requirements is satisfied in the following way:
- Models be encoded in a public standard format
All models in BioModels Database are converted into valid SBML.
- Models must be clearly related to a single reference
Each model is derived from a scientific publication.
- The model must correspond to the biological processes represented in the publication
Models are manually checked to confirm accurate biological representation in the model.
- The simulation results generated from the model must reflect those in the reference publication
Models are manually checked to confirm accurate simulation results are generated from the model. The result of this step is available in the "Curation" tab.
- External data resources annotation
Models elements are well annotated.
What is annotation, and what purpose does it serve in a model?
Once we ensure a model's correctness, we annotate the model with biological information. Model elements can describe a plethora of different entities, such as genes, proteins or metabolites. The usage of free text or non-standard nomenclature to identify those elements is not reliable, as it can introduce ambiguity to the model components, such that subsequent users of a model would struggle to identify the precise entities involved.
Annotation is a process that identifies model elements by linking them to existing data resources. This ensures that users will be able to understand the models and increases the possibility of their reuse. Moreover, annotation is a key element for processes such as data comparison, data integration and data conversion. Finally, it also allows the provision of models accurately through search engines (BioModels Database makes heavy use of annotation when users search for models of interest).
Models in BioModels Database are provided with consistent annotation using unambiguous identifiers. Those identifiers are generated by the MIRIAM Registry and Identifiers.org services. They can be used to reference records from external databases (such as Taxonomy, EMBL-Bank or UniProt), terms from ontologies (such as Gene Ontology, SBO or ChEBI), publications, etc.
For more information about annotation, please refer our general introduction to the annotation of models.
Who annotates the models?
Models can be submitted to BioModels Database already annotated. In this case, any existing annotations encoded using the annotation scheme developed for SBML (introduced since Level 2 Version 2 and supported by numerous tools) will be extracted and recorded in BioModels Database.
Our curation team currently checks existing annotations and adds new ones as necessary. We do encourage authors submitting the models to annotate their models as much as possible by following the guidelines that the curators use during annotation. This will help us accelerate the model curation process.
How are annotations stored in SBML?
The annotation of each model component is stored in the corresponding SBML element using the a scheme initially designed by Nicolas Le Novère and Andrew Finney, and now part of SBML (since Level 2 Version 2). It relies on the use of RDF, Dublin Core, vCard and the BioModels.net qualifiers.
Please refer to the SBML specification for more information about the annotation scheme.
When is a model completely annotated?
Annotating each model component with the most relevant resource records takes great efforts, especially since the number of submitted models has grown rapidly. For example, the 25th release of BioModels Database (18th June 2013) contains 171,432 cross-references (links to external resources contained in the annotations). This number need to be compared with the total number of species (156,665) and relationships (174,986 this number including reactions, rate rules, events and assignment rules) involved in the existing 963 models. Therefore current annotations do not cover all model elements.
Besides manpower limitations, this discrepancy is sometimes due to a lack of adequate or suitable resources for annotation, for instance for molecular entities which are created only for simulation purposes. Moreover, biological data resources are often slightly lagging behind newly generated knowledge, and it is possible that a particular resource does not offer the relevant information at the time the model is annotated. In the case of hierarchical controlled vocabularies, such as Gene Ontology or ChEBI, there is the option to use a term at a higher level of abstraction if the required precise term does not currently exist. Most often, one can always add some information, even if not optimal. Hence, model annotation needs to be, and indeed is, a continuous process.
What are the tools used by the curators of BioModels Database?
Moreover, they also use tools such as Gnuplot and various custom scripts.
What is the non-curated branch of BioModels Database?
The success and rapid growth of BioModels databases has led to a couple of issues, both of which are handled using the non-curated branch of BioModels Database. Firstly, BioModels Database faces an ever-growing, and increasingly complex, number of model submissions which are growing at a pace that exceeds our curation capacity. Secondly, many of the models submitted, while syntactically correct, cannot be curated for various reasons. Hence, the non-curated branch of BioModels Database is a public holding area for models that have yet to be curated, or cannot be curated for various reasons. The models held in the non-curated branch are of four types:
- Models that are perfectly fine, but have yet to be curated.
- Models that are not curatable, since they are not MIRIAM compliant.
- Models that are not currently curatable, since they are not kinetic models (e.g. logical models and Flux Balance Analysis models).
- Models for which SBML supports only part of the description, most of it being in annotation (e.g. spatial models).
Note that all the models in the non-curated branch have nevertheless passed the XML check, SBML syntax check and SBML consistency checks.
What are the tools used to develop and run the software infrastructure behind BioModels Database?
The beta system is based on Jummp, a flexible, open-source model management platform that can be tailored to the needs of other projects or communities. Submissions are hosted using a combination of version control systems such as Git, a relational database such as MySQL or PostgreSQL and a search engine like Lucene.
Can I install my own version of BioModels Database?
The models can be freely reused and the software running BioModels Database itself is an open source project distributed under the GNU General Public License. For more information, please refer to the software infrastructure entry of this FAQ.
All converters are also available under an open source license and can be downloaded from the download section of the BioModels SourceForge repository.
In order to store your own models or just plug part(s) of the BioModels Database infrastructure into your own, you'll need to setup your own local instance of the repository. In order to help you through this procedure, we created the following page: development with BioModels Database source code.
What is the naming and identifier scheme used in BioModels Database?
At the time of submission, a unique submission identifier assigned to the model. It is composed of the character sequence "MODEL" followed by ten digits extracted from the timestamp of submission. When a model is moved to the curated branch, a new model identifier is generated and assigned to it. This identifier is composed of the character sequence "BIOMD" followed by ten digits reflecting the model's position in the branch (for example "BIOMD0000000216" for the 216th model successfully curated).
During the curation process, the curators will also give an appropriate name to each model. This follows the general scheme: AuthorYear - Topic Method. For example, Edelstein1996 - EPSP ACh Event refers to the model BIOMD0000000001.
Both identifiers are unique and permanent, and will never be re-assigned to a different model, even if for some reason a particular model must be retracted from the database. Both identifiers can be used to retrieve the model and quoted in subsequent publications.
Who are the authors of a model?
The authors of a model are the individuals who initially described the model in a peer-reviewed publication.
Who is the submitter of a model?
The submitter of a model is the person who actually submitted the model and therefore requested its addition in BioModels Database. Anybody can submit a model to BioModels Database. Please refer to the submission procedure, if you want to know more.
Who are the encoders of a model?
The encoders of a model are the individuals who actually encoded the model in its present form, based on the published article. These individuals could be the authors of the publication, members of the curation team, or any other scientists who decided to encode the published article.
Why do I get an error message stating some reaction modifiers are not declared?
Some models contain reactions where the kinetics are modified by compounds that are not themselves reactants. These compounds are not consumed or produced by the reaction, and are known as modifiers. This error occurs when some software tools do not explicitly list the modifiers contained in the model reactions. A consistency check has been implemented to identify missing modifier declarations, and is now applied to all models submitted to BioModels Database. This error should not happen any more.