TL;DR
The Reactome pathway knowledgebase has been significantly updated, now encompassing over 13,000 reactions and 2,500 pathways, with a particular focus on SARS-CoV-2 infection and potential therapeutic interventions. This release also introduces new tools for analyzing high-throughput data and integrating "dark" proteins into known biological contexts, enhancing our understanding of complex biological systems.
What Problem Does This Paper Address?
Understanding the intricate network of molecular reactions that drive biological processes in humans is a monumental task. While various resources exist to catalog individual reactions or broader biological pathways, there's a need for a comprehensive, manually curated knowledgebase that focuses on a single species (Homo sapiens) and uses a consistent data model. This allows for a unified view of normal and disease-associated processes, facilitating the discovery of functional relationships within complex biological data. Furthermore, the rapid emergence of new diseases, like COVID-19, necessitates agile methods for annotating viral processes and identifying potential therapeutic targets. The challenge also lies in integrating "understudied" or "dark" proteins, which lack extensive functional annotations, into this existing knowledge framework to uncover their potential roles.
Methods
The Reactome Knowledgebase is built upon a foundation of manual curation by expert scientists, ensuring high-quality and accurate annotations of molecular transformations. For the SARS-CoV-2 infection processes, a novel approach was employed: leveraging the high sequence identity (82%) between SARS-CoV-2 and SARS-CoV-1 to computationally infer pathways, which were then refined and extended with experimental data. This template-based inference method allows for rapid annotation of new viral threats.
New tools have been developed to enhance data analysis and integration:
- ReactomeGSA (Gene Set Analysis): This system supports comparative pathway analysis across multiple experimental datasets, performing differential expression analysis directly at the pathway level. It handles multi-omics and cross-species data by mapping it to Reactome's internal mapping system.
- Reactome IDG Portal: This portal, developed in collaboration with the Illuminating the Druggable Genome (IDG) consortium, aims to place "dark" proteins within the context of Reactome's pathways. It uses data from high-throughput studies and sequence motif conservation to infer potential functions and interactions of these understudied proteins.
Datasets
- Reactome Knowledgebase: The primary dataset is the Reactome Knowledgebase itself, which is continuously updated. The paper details updates from Release 70 (September 2019) to Release 78 (September 2021).
- Ensembl release 104 (May 2021): Used for identifying human protein-coding genes.
- Literature References: Over 34,000 literature references are used to annotate reactions.
- Disease Ontology terms: 623 Disease Ontology terms are used to tag disease-specific reactions.
- Drug Information: Data from the Guide to Pharmacology 'Coronavirus information' resource and recent publications are used to catalog potential drug targets.
- Public Datasets: ReactomeGSA can directly integrate public datasets from ExpressionAtlas and Single Cell Expression Atlas.
- Target Central Resource Database (TCRD): Used for overlaying protein expression values and protein-protein interactions.
Key Insights and Ideas
- Agile COVID-19 Annotation: The ability to rapidly annotate viral infection pathways by leveraging existing knowledge of related viruses (SARS-CoV-1) is a significant advancement for responding to emerging infectious diseases.
- Bridging the "Dark Proteome" Gap: The IDG portal offers a crucial step towards understanding the function of understudied proteins by placing them within the context of known biological pathways and interactions. This can guide future experimental research.
- Integrated Multi-omics Analysis: ReactomeGSA's ability to perform pathway-level differential expression analysis across diverse 'omics data types and species streamlines the interpretation of complex biological experiments.
- Drug Repurposing for COVID-19: The systematic cataloging of potential drug targets and their integration into viral infection pathways provides a valuable resource for identifying and evaluating repurposed therapeutics.
Key Results
- Reactome Content Growth: Release 78 (October 2021) contains 10,726 human proteins (52.5% of predicted genes), involved in 13,890 reactions, grouped into 2,546 pathways and 28 superpathways.
- Disease Annotations: 4,603 variant proteins and 1,544 disease-specific reactions are annotated.
- Drug Annotations: 507 drugs are annotated for their modulating effects on normal and disease processes.
- SARS-CoV-2 Pathway Coverage: 128 reactions comprise the SARS-CoV-2 infection process, with 116 having associated SARS-CoV-2-specific data.
- New Tools: ReactomeGSA and the IDG portal are now available, offering enhanced analytical capabilities.
- Visual Enhancements: The number of textbook-style pathway diagrams has increased from 91 to 150, and the biomolecular icon library has grown from 1,350 to 2,040.
Limitations and Future Work
- "Dark" Protein Annotation: While the IDG portal aims to address the "dark proteome," the inherent lack of experimental data for these proteins means that inferred functions are hypotheses that require experimental validation.
- Continued Curation: The Reactome Knowledgebase is a living resource that requires continuous updates as new research emerges, particularly for rapidly evolving areas like infectious diseases.
- Integration with Other Resources: While Reactome integrates data from various sources, further efforts to harmonize and link with other biological databases could enhance its utility.
- Experimental Validation: The paper highlights the reliance on computational inference for some annotations, underscoring the ongoing need for experimental validation to confirm these predictions.
Summary updated