Open Targets Platform

Open Targets Platform integrates diverse public datasets to systematically identify and prioritize therapeutic drug targets by scoring and ranking target-disease associations.

Key Features:

• Evidence Integration: Aggregates genetics, genomics, transcriptomics, drugs, animal models, and scientific literature to support target-disease associations, including genome-wide CRISPR knockout screens in cancer models (Project Score) and GWAS/UK BioBank statistical genetic evidence.
• Enhanced Evidence Scoring: Applies an evolved scoring framework that incorporates prioritized targets, curated target tractability and safety annotations, and evaluation of post-marketing adverse drug reactions.
• Data Harmonization and Safety/Tractability Assessments: Implements continuous data integration and harmonization with assessments of target safety and tractability using genetic constraint, PROTACtability, and AlphaFold structure predictions.
• Machine Learning Extraction: Uses machine learning to extract knowledge from scientific literature, clinical trials, and drug labels.
• Data Formats: Produces datasets in JSON and Parquet formats.

Scientific Applications:

• Target identification and prioritization: Systematically identifies and ranks therapeutic targets for drug discovery using integrated multi-evidence scoring.
• Genetic association interpretation: Supports interpretation of GWAS/UK BioBank statistical genetic evidence for linking targets to diseases.
• Functional genomics and cancer target discovery: Leverages genome-wide CRISPR knockout screens (Project Score) in cancer models to inform target selection.
• Safety and tractability evaluation: Assesses target safety and tractability via genetic constraint, PROTACtability, AlphaFold predictions, and post-marketing adverse reaction data.
• Knowledge mining: Extracts and synthesizes evidence from literature, clinical trials, and drug labels to augment target-disease associations.

Methodology:

Integrates genetics, genomics, transcriptomics, drug, animal model, and literature datasets; scores and ranks target-disease associations; incorporates genome-wide CRISPR knockout screens (Project Score) and GWAS/UK BioBank statistical genetic evidence; includes prioritized target lists and curated tractability/safety annotations; evaluates post-marketing adverse drug reactions; assesses targets using genetic constraint, PROTACtability, and AlphaFold structure predictions; and applies machine learning to extract knowledge from literature, clinical trials, and drug labels.