MolData

MolData curates bioassay results and annotates each record with specific disease and molecular target labels to enable machine-learning-driven drug discovery and drug-repurposing analyses.

Key Features:

• Targeted Data Curation: MolData assigns each bioassay result to specific diseases and targets, producing labeled datasets for downstream analysis.
• Extensive Coverage: The dataset encompasses 30 unique categories of diseases and targets, providing broad biomedical coverage.
• Facilitation of Machine Learning: Explicit target and disease annotations support training and evaluation of machine learning models and provide multitask-learning benchmarks with over 30 models per category.
• Drug Repurposing Insights: Correlation analysis of annotated assay results identifies potential drug-repurposing opportunities across diseases including cancer, metabolic disorders, and infectious diseases.

Scientific Applications:

• Machine learning training and evaluation: Structured, labeled bioassay data for training and benchmarking predictive models.
• Multitask learning benchmarking: Benchmarks for multitask models across the 30 disease and target categories.
• Identification of drug candidates: Discovery of potential molecular candidates through analysis of assay annotations and model predictions.
• Drug repurposing analysis: Identification of repurposing opportunities across cancer, metabolic disorders, and infectious diseases via correlation analysis.
• Academic and pharmaceutical research: Use in academic research and pharmaceutical development for computational drug discovery workflows.

Methodology:

Assignment of bioassay results to disease and target labels; organization of data with explicit target and disease annotations for machine learning; provision of multitask-learning benchmarks with over 30 models trained per category; correlation analysis of annotated assay results for drug-repurposing insight.