molfeat

Convert SMILES strings into machine-learning feature vectors using 100+ featurizers.

16|7|Updated Nov 20, 2025
One-click install
npx skills add https://github.com/jackspace/ClaudeSkillz --skill molfeat
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: molfeat
Source: https://github.com/jackspace/ClaudeSkillz/tree/main/skills/scientific-pkg-molfeat
Command: npx skills add https://github.com/jackspace/ClaudeSkillz --skill molfeat

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

MolFeat unifies 100+ featurizers (ECFP, MACCS, descriptors, pretrained models like ChemBERTa) and provides ready-to-use pipelines to convert SMILES to machine-learning features, speeding up model development and deployment in cheminformatics.

Core Features & Use Cases

  • Single API for many featurizers: Mix fingerprints, descriptors, and embeddings.
  • Batch transformers: Scikit-learn compatible transformers for pipelines.
  • Performance & reproducibility: Parallel processing, caching, and state saving.
  • Common workflows: QSAR, virtual screening, similarity search, deep learning embeddings.
  • Use cases include traditional ML with fingerprints, interpretable descriptors, and transformer-based embeddings.

Quick Start

Featurize a small set of SMILES with FPCalculator and MoleculeTransformer, then train a simple model.

Frequently Asked Questions about molfeat

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert SMILES strings into machine learning features?

Featurize SMILES strings into ML-ready vectors using molfeat's 100+ featurizers including ECFP, MACCS, descriptors, and pretrained models like ChemBERTa. Accept both SMILES and RDKit Mol objects, output numerical feature vectors for direct model training.

Can I combine multiple fingerprints and descriptors in a single pipeline?

Yes, molfeat enables unified pipelines that mix fingerprints, descriptors, and embeddings into one representation. Use batch transformers compatible with scikit-learn to combine featurizers and feed the result directly into ML workflows.

What featurization methods work for QSAR and virtual screening?

Molfeat provides ECFP, MACCS, MAP4, molecular descriptors, and pretrained transformer embeddings—all applicable to QSAR modeling, virtual screening, similarity searching, and chemical-space analysis at scale with parallelization and caching.

How do I handle large-scale molecular featurization efficiently?

Molfeat implements parallel processing, error handling, and caching to featurize large molecular datasets. Batch transformers and state-saving enable reproducible, scaled pipelines for production cheminformatics workflows.

Can I use pretrained models for molecular embeddings?

Yes, molfeat exposes pretrained transformers such as ChemBERTa alongside traditional fingerprints and descriptors. Pretrained transformers generate high-quality embeddings for deep-learning pipelines and representation learning.

Do I need RDKit to use molfeat?

Molfeat accepts both SMILES strings and RDKit Mol objects as input. The unified API handles molecular structure conversion, so you can work with either format depending on your workflow.