Virtual Cell at UConn



BioPAX at Virtual Cell

The primary challenge for mapping of BioPAX data to SBML is to find out, which of all instances of a physical entity (called Physical Entity Participants in BioPAX) are mapped to the same species Type in SBML. After Physical Entities are mapped to speciesTypes, mapping to species is based on location. Below we describe some of reasoning applied by SybReaM, a Systems Biology Reasoning Engine for Analysis and Modeling, that provides such a mapping in Sybil.

We call a physical entity in BioPAX united, if in all interactions in which the physical entity participates at a given location, all of it can participate. Otherwise, a physical entity is called divided. If a physical entity is divided, then a division is a part of the physical entity for which there is an interaction such that all of the division, and no part outside of the division, can participate in this interaction. The divisions will be mapped to SBML species Note that different divisions are not necessarily mutually disjoint, and their union is not necessarily equivalent to the physical entity.

A variety of assumptions are used to determine all divisions of a physical entity, and to determine which interactions involve which division. A typical assumption is that all physical entities of BioPAX class small molecule are united. Another typical assumption is that two physical entity participants appearing on both sides of a conversion, referring to the same physical entity, belong to different divisions.
Yet another assumption is that physical entity participants with incompatible lists of sequence features belong to different divisions (more precisely, we are talking here of sequence participants). This assumption may not always hold, since lists of sequence features may not be exhaustive. It is somewhat stronger in cases where one physical entity participant has no sequence features listed. Another (very weak and not always correct) assumption is if the name of the physical entity participant contains phrases like phospho which do not appear in the name of the physical entity, then physical entity is divided. On the other hand, an assumption may be made that if two conversions are consecutive pathway steps, and the only connection between them is through a physical entity to which they both refer through one physical entity participant each, then these two physical entity participants belong to the same divisions. In the worst case, there may be as man divisions as there are interactions, but usually, their number is less, meaning, the same division can participate completely and exclusively in multiple interactions.