Does this skill support Graph Neural Networks?

Yes, it includes specialized featurizers and model architectures for GCNs, Graph Attention Networks (GAT), Message Passing Neural Networks (MPNN), and AttentiveFP.

Can I use pre-trained models with this skill?

Absolutely. It supports transfer learning workflows using pre-trained models like ChemBERTa and GROVER, which are particularly effective for improving performance on small, high-value datasets.

What is DeepChem used for?

DeepChem is a comprehensive Python library designed to apply machine learning to chemistry, materials science, and biology, specifically for tasks like molecular property prediction and drug discovery.

Why should I use ScaffoldSplitter instead of RandomSplitter?

ScaffoldSplitter is critical in chemistry tasks because it ensures molecules with similar structural cores are in the same set, preventing data leakage and providing a more realistic assessment of how a model performs on truly novel molecules.

DeepChem Molecular ML

Name: DeepChem Molecular ML
Author: Yezez9

byYezez9

0•

データサイエンスとML

Simplifies molecular machine learning workflows for property prediction, drug discovery, and chemical data analysis using specialized featurizers and pre-trained models.

DeepChem for Claude Code provides specialized capabilities for chemical and biological data science, enabling developers to build sophisticated models for molecular property prediction, ADMET analysis, and drug discovery. It streamlines the complex process of loading SMILES strings or SDF files, applying advanced molecular featurizers like circular fingerprints or graph neural networks, and utilizing standardized MoleculeNet benchmarks. Whether you are performing scaffold splitting to prevent data leakage or fine-tuning pre-trained transformers like ChemBERTa, this skill offers a robust, research-ready framework for scientific machine learning tasks.

主な機能

010 GitHub stars

02Seamless access to 30+ MoleculeNet benchmark datasets like Tox21, BBBP, and Delaney

03Comprehensive molecular featurization including fingerprints, descriptors, and graph representations

04Specialized molecular data splitting methods including Scaffold and Butina splitting

05Support for Graph Neural Networks such as GCN, GAT, MPNN, and AttentiveFP

06Integration with pre-trained models like ChemBERTa and GROVER for transfer learning

ユースケース

01Predicting molecular toxicity and ADMET properties for drug discovery pipelines

02Benchmarking new machine learning architectures on standardized chemical datasets

03Analyzing protein sequences and crystal structure properties for materials science

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yezez9/research-agent deepchem

For use in Claude.ai and ChatGPT