Software infrastructure to support the standard of model curation and annotation MIRIAM

Systems Biology recently emerged as on of main field in life sciences. A crucial component of this discipline is the development of quantitative models to simulate the biological processes. Scientists need to store and exchange those models to leverage on each-other expertise and production. A prerequisite is the standardisation of model's syntax and content. In collaboration with the community of modellers and developers, we developed a standard, MIRIAM, that precises what mandatory information authors have to add to a quantitative model in order to facilitate its understanding by other scientists. MIRIAM is made up of two parts: 1) a set of rules controlling the correspondance between the model and its description in the scientific literature; 2) A set of annotation describing the generation of the model, and documenting each of its elements. For MIRIAM to be effectively adopted, we need to provide the computing resources necessary to its implementation. We will build a database of officialy endorsed data resources. This database will provide for each data resource all the details necessary to the implementation of the standard. We will develop a simple interface to facilitate the automatic access to the database. Software libraries will be provided to implement the access to the interface and to facilitate the development of third party software. In collaboration with the California Institute of Technology and the Bioengineering Institute of Auckland we will work to expand the support of the standard MIRIAM by the modelling languages SBML and CellML. This will allow a significant fraction of modelling tools to produce MIRIAM-compliant models. The endorsement of public standards by scientists, developers and publishers will be a crucial step toward the effective storage, exchange and reuse of quantitative models in Biology, one of the cornerstone of Systems Biology. The wide acceptance of MIRIAM will only be achieved through the availability of easy to use supporting softwares and resources such as those presented in this project.

To improve the exchange and interpretation of quantitative models, we developed the Minimal Information Requested In the Annotation of Models (MIRIAM), a standard that summarises the mandatory information authors have to add to a model. For MIRIAM to be effectively adopted, we need to provide the computing resources necessary to its implementation. 1) We will build a relational database (using MySQL) of officialy endorsed data resources, with all the necessary information to access and retrieve information from them. This database will provide, for each agreed-upon data resource, all the details necessary to the implementation of MIRIAM: standard URIs, corresponding physical URLs, names of resource, description of the syntax to build identifiers and hyperlinks etc. 2) In order to facilitate the programmatic access to the database, we will develop a simple Application Programming Interface. This API will be available through WebServices. The server side will be developed using Java Servlet and the Axis engine of the Apache project. We will provide libraries implementing this access to facilitate the development of MIRIAM clients. 3) In collaboration with the SBML-team at the California Institute of Technology, and the CellML team at the Bioengineering Institute of Auckland, we will work to expand the support of MIRIAM annotation in SBML and CellML. In particular, we will expand the software library libSBML to support MIRIAM annotation. This will allow all the software using libSBML (which means a significant fraction of SBML-supporting tools) to produce MIRIAM-compliant models.