pytdc-therapeutics-data-commons

What problem does it solve?

This Skill provides easy access to a comprehensive platform of AI-ready datasets and benchmarks specifically designed for drug discovery, streamlining model development and evaluation.

Core Features & Use Cases

• Access Diverse Datasets: Load curated datasets for ADME, toxicity, drug-target interactions (DTI), and more.
• Standardized Evaluation: Utilize built-in split methods (scaffold, cold splits) and evaluation metrics for robust benchmarking.
• Molecule Generation: Access training sets and oracles for molecular design and retrosynthesis tasks.
• Use Case: Train a machine learning model to predict drug toxicity by loading the ClinTox dataset, applying a scaffold split, and evaluating performance using ROC-AUC.

Quick Start

Load the Caco2_Wang ADME dataset and get a scaffold split with seed 42.