Molfeat is a Python library and Claude Skill that provides a unified interface for over 100 molecular featurization methods used in machine learning tasks.

Do I need to install additional dependencies for deep learning?

Specific featurizers like GNNs or Transformers require optional dependencies such as 'molfeat[dgl]' or 'molfeat[transformer]' to be installed via pip.

Which molecular representations does Molfeat support?

It supports traditional fingerprints like ECFP and MACCS, 2D/3D descriptors like Mordred, and deep learning embeddings like ChemBERTa and Graphormer.

Can I process large chemical datasets efficiently?

Absolutely. Molfeat features built-in parallel processing and batching capabilities to handle large-scale chemical libraries and high-throughput screening data.

Is it compatible with scikit-learn?

Yes, Molfeat includes MoleculeTransformer classes that are fully compatible with scikit-learn pipelines, allowing for seamless integration into existing ML models.

Molfeat Molecular Featurization

Name: Molfeat Molecular Featurization
Author: K-Dense-AI

byK-Dense-AI

•

2,188

数据科学与机器学习

Converts chemical structures into high-dimensional numerical representations for molecular machine learning and QSAR modeling.

关于

Molfeat is a comprehensive Python library and specialized Claude Skill designed to streamline the featurization of chemical structures for machine learning. It unifies over 100 pre-trained embeddings and hand-crafted featurizers, enabling researchers and developers to easily convert SMILES strings or RDKit molecules into machine-ready feature vectors. Whether you are building QSAR models, performing virtual screening, or conducting similarity searches, Molfeat provides a standardized interface with built-in parallelization and scikit-learn compatibility to accelerate molecular data processing workflows.

主要功能

Access to 100+ pre-trained embeddings and traditional chemical fingerprints
Support for deep learning models like ChemBERTa and Graph Neural Networks
Seamless conversion from SMILES strings to molecular feature vectors
Unified interface for calculators and scikit-learn compatible transformers
High-performance batch processing with automatic parallelization
2,188 GitHub stars

使用场景

Conducting large-scale virtual screening for drug discovery hits
Performing chemical space analysis and structural similarity searching
Building Quantitative Structure-Activity Relationship (QSAR) models

关于

主要功能

Access to 100+ pre-trained embeddings and traditional chemical fingerprints
Support for deep learning models like ChemBERTa and Graph Neural Networks
Seamless conversion from SMILES strings to molecular feature vectors
Unified interface for calculators and scikit-learn compatible transformers
High-performance batch processing with automatic parallelization
2,188 GitHub stars

使用场景

Conducting large-scale virtual screening for drug discovery hits
Performing chemical space analysis and structural similarity searching
Building Quantitative Structure-Activity Relationship (QSAR) models