Reactome and the Gene Ontology: Digital pathway convergence for core data resources

点击翻译按钮获取中文摘要

The Gene Ontology (GO) knowledgebase is the world’s largest source of information on the functions of genes, and is a foundation for computational analysis in biomedical research. Reac- tome integrates information about the interconnected biological pathways that describe human biology at a molecular level. The proposed work will extend a collaboration between the GO and Reactome projects to better integrate content currently siloed in the two resources and several related reference resources to provide a more fully interoperable open access resource for the biomedical research community. GO provides widely-used tools for associating an individual gene product with a cellular compo- nent of which it is a part, a molecular function it mediates, and a biological process in which it participates. Annotations of biological reactions in the Reactome Knowledgebase integrate many such atomic GO assertions concerning gene products as they interact in biological pro- cesses, information of great potential value to researchers who use GO annotations, but hard to use because the two are formulated differently. The recently developed GO Causal Activity Model (GO-CAM) formalism, which extends atomic GO annotation to capture and integrate en- tire biological processes, can form the basis of integration between GO and Reactome's path- way models, increasing their utility. We propose to develop an integrated pathway/ontology resource by developing tools to effi- ciently extract GO-CAM model content and further integrate GO-CAM / Reactome models with key reference resources such as Rhea (a database of expert-curated biochemical reactions) and the Protein Ontology (PRO), an ontological representation of protein-related entities. Specifically, we will 1) generate GO-CAM models for every Reactome pathway that represents a normal human biological process; 2) coordinate curation and QA between Reactome and GO; 3) support data exchange among resources and cross-validate the mappings; and 4) extend this work to support efficient, partly automated creation of pathway resources for model organisms. The end result will be an open-source, integrated set of tools and data that accurately represent human and model organism biological processes, supporting diverse data mining, analysis, and modeling activities. These tools and data will be incorporated into GO, Reactome, PRO, and participating model organism resources, supporting long-term alignment of key community re- sources after the completion of this project.