Therapeutics Data Commons. AI-ready drug discovery datasets (ADME, toxicity, DTI), benchmarks, scaffold splits, molecular oracles, for therapeutic ML and pharmacological prediction.
8.7
Rating
0
Installs
Machine Learning
Category
Excellent skill that provides comprehensive access to PyTDC's drug discovery datasets and tools. The SKILL.md is exceptionally well-structured with clear organization across single-instance prediction, multi-instance prediction, and generation tasks. The description coverage is outstanding—a CLI agent can easily understand when and how to use the skill for ADME, toxicity, DTI, molecular generation, and benchmarking tasks. Task knowledge is thorough with code examples for all major workflows, and the skill appropriately references external files for detailed catalogs (datasets.md, oracles.md, utilities.md) and scripts, keeping the main file focused yet complete. The structure is exemplary: concise overview, clear categorization, quick-start patterns, and logical progression from basic to advanced features. Novelty is strong—while PyTDC is a wrapper library, it provides specialized domain knowledge for pharmaceutical ML that would require significant prompt engineering for a CLI agent to navigate correctly (knowing proper split methods like scaffold/cold splits, understanding ADMET properties, using oracles for molecular optimization). This meaningfully reduces token cost and error risk for complex drug discovery workflows. Minor deduction on novelty only because it's primarily a data access layer rather than implementing novel algorithms, but the domain specialization and workflow orchestration provide clear value.
Loading SKILL.md…