What is PyTDC used for in Claude Code?

PyTDC is used to retrieve, split, and evaluate datasets for therapeutic machine learning tasks like predicting drug toxicity, binding affinity, and molecular property optimization.

Can I use PyTDC for molecular generation tasks?

Yes, the skill provides access to MolGen datasets and 17+ oracle functions to evaluate and optimize the properties of generated molecules.

What data splitting methods does the PyTDC skill support?

It supports advanced strategies including scaffold splits for chemical diversity, cold-drug/cold-target splits for zero-shot generalization, and standard random or temporal splits.

How do I install the necessary dependencies for this skill?

You can install the core library using 'uv pip install PyTDC', which automatically handles base dependencies like numpy, pandas, and scikit-learn.

PyTDC Drug Discovery

Name: PyTDC Drug Discovery
Author: K-Dense-AI

byK-Dense-AI

•

2,188

Ciencia de Datos y ML

Accesses AI-ready Therapeutics Data Commons (TDC) datasets and benchmarks for drug discovery and pharmaceutical machine learning.

Acerca de

PyTDC (Therapeutics Data Commons) is a specialized skill that empowers Claude to assist with the entire therapeutics pipeline, from early-stage drug discovery to clinical trial outcome prediction. It provides a standardized interface for curated datasets covering ADMET properties, drug-target interactions, and molecular generation. The skill includes advanced data processing utilities such as scaffold and cold-start splits, molecular oracles for property optimization, and rigorous benchmark groups to ensure scientifically valid machine learning evaluations in the pharmacological domain.

Características Principales

Standardized access to ADME, Toxicity, and Drug-Target Interaction (DTI) datasets
Advanced data splitting strategies including scaffold, cold-start, and temporal splits
Comprehensive benchmark groups for systematic 5-seed model evaluation
Integrated molecular oracles for evaluating and optimizing chemical properties
2,188 GitHub stars
Chemical utility functions for SMILES/SELFIES conversion and molecular filtering

Casos de Uso

Predicting pharmacokinetic properties and toxicity of novel drug candidates
Benchmarking machine learning models on standardized pharmaceutical datasets
Generating and optimizing molecules with desired bioactivity using oracle functions

Acerca de

Características Principales

Standardized access to ADME, Toxicity, and Drug-Target Interaction (DTI) datasets
Advanced data splitting strategies including scaffold, cold-start, and temporal splits
Comprehensive benchmark groups for systematic 5-seed model evaluation
Integrated molecular oracles for evaluating and optimizing chemical properties
2,188 GitHub stars
Chemical utility functions for SMILES/SELFIES conversion and molecular filtering

Casos de Uso

Predicting pharmacokinetic properties and toxicity of novel drug candidates
Benchmarking machine learning models on standardized pharmaceutical datasets
Generating and optimizing molecules with desired bioactivity using oracle functions