What datasets can I access with this skill?

You can access over 30 curated MoleculeNet benchmarks including Tox21, BBBP, Delaney (Solubility), and HIV datasets for standardized model evaluation.

Does this skill handle molecular structure files?

Yes, it includes specialized loaders for SMILES strings, SDF files, FASTA protein sequences, and JSON-formatted chemical data.

Why should I use ScaffoldSplitter instead of RandomSplitter?

ScaffoldSplitter is essential for chemistry because it ensures that molecules with similar structural cores are not split between training and test sets, providing a more realistic measure of model generalization.

Can this skill predict crystal properties for materials science?

Yes, DeepChem supports materials-specific models like CGCNN and MEGNet for predicting properties such as bandgap and formation energy in crystals.

DeepChem Molecular Machine Learning

Name: DeepChem Molecular Machine Learning
Author: K-Dense-AI

byK-Dense-AI

•

3,718

•

数据科学与机器学习

Empowers molecular machine learning and drug discovery through advanced chemical featurization, property prediction, and Graph Neural Networks.

This skill integrates the DeepChem library into Claude's workflow, enabling sophisticated chemical and biological data analysis. It covers the entire lifecycle of molecular machine learning, from loading SMILES strings and SDF files to implementing Graph Neural Networks (GNNs) and utilizing MoleculeNet benchmarks. By providing domain-specific patterns like scaffold splitting to prevent data leakage and transfer learning with models like ChemBERTa, it helps researchers and developers build robust models for ADMET prediction, toxicity screening, and materials design.

主要功能

01Seamless access to 30+ MoleculeNet benchmark datasets

023,718 GitHub stars

03Specialized chemical data splitters including Scaffold and Butina methods

04Comprehensive Graph Neural Network support (GCN, GAT, MPNN, AttentiveFP)

05Transfer learning workflows for pretrained molecular transformers

06Automated molecular featurization (Fingerprints, Graph representations, Descriptors)

使用场景

01Implementing graph-based models for complex molecular property analysis

02Fine-tuning pretrained chemical transformers on small biological datasets

03Building toxicity and solubility prediction models for drug discovery

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add k-dense-ai/claude-scientific-skills deepchem

For use in Claude.ai and ChatGPT

Download Skill