Paxtools

Paxtools is a Java library, API, and command-line toolkit that provides the reference object model for BioPAX (Biological Pathway Exchange), the RDF/OWL standard for representing molecular and cellular pathways (format:biopax-l3, format:owl). It was developed at the Computational Biology Center of Memorial Sloan Kettering Cancer Center together with the Bader Lab at the University of Toronto, and it is the software foundation behind the BioPAX ecosystem and the Pathway Commons database. The project is Maven-based and modular, with a compact core model plus optional modules for I/O, conversion, graph queries, and a Pathway Commons web-service client.

Its central abstraction is an in-memory BioPAX Model: a container of typed Java beans that mirror the BioPAX class hierarchy (physical entities, interactions, conversions, controls, pathways, and their controlled vocabularies and cross-references). Because BioPAX is expressed in OWL, Paxtools implements OWL property semantics that plain Java lacks, automatically maintaining inverse links, symmetric and transitive relations, and domain/range consistency as the model is edited. A reflection-based PropertyEditor layer lets code traverse and mutate the model generically without hard-coding property names, which underpins its validation and query utilities. Paxtools handles BioPAX Levels 2 and 3 and can upgrade older models to Level 3; Level 1 support was removed as of version 5.

For the format-conversion and graph-data world, Paxtools is primarily an exporter and simplifier. It reads and writes native BioPAX OWL and converts to Simple Interaction Format (format:sif), a reduction that collapses detailed mechanistic states into binary edges suitable for network analysis and visualization in tools like tool:cytoscape. It also exports to SBGN-ML (format:sbgn-ml-0.3) via tool:libsbgn and to GSEA gene sets, and it converts PSI-MI molecular-interaction data (format:psi-mi-xml-3.0) into BioPAX Level 3. Its graph algorithms answer biological questions such as paths-between, neighborhood, and common upstream/downstream searches over pathway networks.

Paxtools is scriptable from Java or Jython and exposes common operations through a console command. It is wrapped for R as tool:paxtoolsr and for network export as tool:pax2graphml. Its main strengths are standards-correct BioPAX handling, semantic integrity, and mature conversion rules; its limitations are that it is BioPAX-specific rather than a general graph library, SIF conversion is inherently lossy, and effective use assumes familiarity with the fairly intricate BioPAX schema.