The goal of our research project is to develop a platform for the Virtual Cell which would allow the integration of data in BioPAX. To accomplish this, we developed Sybil, which ports and links between BioPAX and SBML, which ultimately allows for import of BioPAX data to the Virtual Cell, because the Virtual Cell understands SBML. But Sybil is a useful tool for any SBML and BioPAX integration, in other words for any one who wants to use BioPAX with an SBML-capable platform or vice versa. To perform its function, Sybil uses an RDF/OWL-based ontology called SBPAX to link between SBML and BioPAX, and a automatic reasoner called SYBREAM.
The Virtual Cell is a simulation platform for cellular reaction networks. The Virtual Cell has been being developed by the Virtual Cell team at the Center for Cell Analysis and Modeling (CCAM) at the Department for Cell Biology at the University of Connecticut Health Center. It includes now a wide variety of tools for simulation and evaluation. Researchers from all over the world can access the VCell on the web by registering as users and downloading the VCell client. Users can then design and store models with the client and submit models for simulation to the VCell servers. The VCell has an XML-based data format called Virtual Cell Markup Language (VCML), which has great similarities with SBML, and the VCell performs conversions between VCML and SBML.
Systems Biology Markup Language (SBML) is an XML-based standard for describing kinetic models of cellular metabolic networks and pathways. Developed by a workgroup, it has evolved into a mature standard consisting of a core and a number of extensions. The basic model in SBML consists of reactions involving a kinetic law and a number of species representing reactants, products and calaysts, typically representing amounts of a certain susbstance in a certain place. A simulator can, for example, use the kinetic laws to formulate and solve differential equations for the species as functions over time. Since SBML is a standard, it allows models to be stored in a database for download by users.
Biological Pathway Exchange (BioPAX) is a standard developed by a different community, which is also designed to represent cellular metabolic pathways, and based on Semantic Web standards such as RDF and OWL. Unlike SBML, BioPAX does not support kinetic modeling, but as you would expect from a Semantic Web oriented format, it has a richer semantic structure, by defining a hierarchy of terms to describe objects and relationships. In BioPAX, there are physical entities and interactions. Interactions can be either conversions, which means they are interactions between physical entities, or they can be controls (catalysis or modulation), which means they are interactions between a physical entity and another interaction. BioPAX also supports linking objects to a variety of references, such as publications, websites or database entries. A number of databases offer pathway data in BioPAX, totalling tens of thousands of interactions.
Since both SBML and BioPAX talk about cellular metabolic networks in terms which seem similar, and both come with respective communities possessing large amounts of data, there is an obvious benefit in making it possible for both communities to exchange data. For example, building a model for the VCell using data in BioPAX. In fact, some tools exist for conversion.
But these tools assume one-to-one relationship, which is in general not a good assumption. Differences arise between SBMl and BioPAX for various reasons: BioPAX attempts to describe objective reality, while SBML attempts to describe models, which contain subjective choices of modelers. BioPAX attempts to classify physical entities according to a universal scheme, while SBML is much more flexible in what is allowed as a species.
Systems Biology Linker (Sybil) intends to work even where this assumption is not valid. Since conversion alone is not enough, we are persuing a more powerful approach. Instead of converting BioPAX to SBML and then throwing away the original data, we keep try to keep both SBML and BioPAX data and link a common subset of both together.
Linking requires a language powerful enough to express everything which can be expressed by a subset of SBML and a subset of BioPAX. We have developed such a language in the form of a RDF/OWL-based ontology we call Systems Biology Pathway Exchange (SBPAX).
In essence, we build a joint repository containing SBML, BioPAX and SBPAX components. Building these components requires intelligence. As much as possible, this intelligence is provided by the Systems Biology Reasoning Engine for Analysis and Modeling (SYBREAM).
I also have some links to my research on my homepage at the Technology Centers for Networks and Pathways (TCNP) Wiki, and there you can also find pages about Sybil and SBPAX.