Reactome is an expert-authored, peer-reviewed knowledgebase of human biological pathways and processes. It serves as a data mining resource and electronic textbook, containing 2975 human proteins, 2907 reactions, and 4455 literature citations. The current release includes a new entity-level pathway viewer and improved search and data mining tools for pathway data visualization and analysis of user-supplied high-throughput datasets. Reactome has expanded its utility to model organism communities through improved orthology prediction methods, allowing pathway inference for 22 species and collaborations to create manually curated Reactome pathway datasets for species such as Arabidopsis, rice, Drosophila, and chicken. The current release of Reactome (version 26, September 2008) covers approximately 12.5% of 20,000 curated UniProt human proteins, a 2.7-fold increase over the last three years. Forty-six major domains of human biology are annotated, including apoptosis, HIV and influenza life cycles, DNA replication, transcription, hemostasis, and carbohydrate metabolism. Normal functions of 1005 proteins associated with OMIM disease phenotypes are also annotated. Improved orthology prediction methods using the OrthoMCL clustering procedure have been implemented to identify orthologs of curated human proteins in 22 evolutionarily divergent species. These changes have improved the success rate for electronic inference without affecting accuracy. Improved tools for analysis of large-scale datasets, data mining, and modeling include SkyPainter, which allows users to visualize functional relationships among genes identified in large-scale experiments. The Reactome BioMart tool facilitates data mining, cross-database analysis, and large-scale analysis of gene function. Collaborations have been established to create model organism Reactome projects for Arabidopsis, rice, Drosophila, and chicken. The Arabidopsis Reactome contains seven manually curated pathways and 311 inferred pathways from KEGG and MetaCyc. Drosophila and chicken reactions are being curated in collaboration with FlyBase and the Gallus Knowledge Extraction and Annotation project. The life cycle of the Influenza virus has been curated and is publicly accessible as part of Reactome. Reactome annotations of complexes and physiological states of proteins are being applied to build a hierarchy of proteins and their functional forms in collaboration with the Protein Ontology group. Reactome data is increasingly integrated and accessible through improved online documentation. Reactome pathway annotations are included in the Pathway Interaction Database. Reactome's online documentation is available as WIKI pages. Reactome content can be exported in SBML, Protégé, Cytoscape, and BioPax formats. A SOAP-based Web Services API is available for accessing Reactome data. A long-term objective of the Reactome project is to provide users with intuitive graphical representations of pathways and reactions. A beta version of an entity-level pathway visualization tool hasReactome is an expert-authored, peer-reviewed knowledgebase of human biological pathways and processes. It serves as a data mining resource and electronic textbook, containing 2975 human proteins, 2907 reactions, and 4455 literature citations. The current release includes a new entity-level pathway viewer and improved search and data mining tools for pathway data visualization and analysis of user-supplied high-throughput datasets. Reactome has expanded its utility to model organism communities through improved orthology prediction methods, allowing pathway inference for 22 species and collaborations to create manually curated Reactome pathway datasets for species such as Arabidopsis, rice, Drosophila, and chicken. The current release of Reactome (version 26, September 2008) covers approximately 12.5% of 20,000 curated UniProt human proteins, a 2.7-fold increase over the last three years. Forty-six major domains of human biology are annotated, including apoptosis, HIV and influenza life cycles, DNA replication, transcription, hemostasis, and carbohydrate metabolism. Normal functions of 1005 proteins associated with OMIM disease phenotypes are also annotated. Improved orthology prediction methods using the OrthoMCL clustering procedure have been implemented to identify orthologs of curated human proteins in 22 evolutionarily divergent species. These changes have improved the success rate for electronic inference without affecting accuracy. Improved tools for analysis of large-scale datasets, data mining, and modeling include SkyPainter, which allows users to visualize functional relationships among genes identified in large-scale experiments. The Reactome BioMart tool facilitates data mining, cross-database analysis, and large-scale analysis of gene function. Collaborations have been established to create model organism Reactome projects for Arabidopsis, rice, Drosophila, and chicken. The Arabidopsis Reactome contains seven manually curated pathways and 311 inferred pathways from KEGG and MetaCyc. Drosophila and chicken reactions are being curated in collaboration with FlyBase and the Gallus Knowledge Extraction and Annotation project. The life cycle of the Influenza virus has been curated and is publicly accessible as part of Reactome. Reactome annotations of complexes and physiological states of proteins are being applied to build a hierarchy of proteins and their functional forms in collaboration with the Protein Ontology group. Reactome data is increasingly integrated and accessible through improved online documentation. Reactome pathway annotations are included in the Pathway Interaction Database. Reactome's online documentation is available as WIKI pages. Reactome content can be exported in SBML, Protégé, Cytoscape, and BioPax formats. A SOAP-based Web Services API is available for accessing Reactome data. A long-term objective of the Reactome project is to provide users with intuitive graphical representations of pathways and reactions. A beta version of an entity-level pathway visualization tool has