What molecular formats does Molfeat support?

Molfeat primarily supports SMILES strings and RDKit molecule objects as inputs for its calculators and transformers.

Is Molfeat compatible with scikit-learn?

Absolutely. Molfeat's MoleculeTransformer is designed to be fully compatible with scikit-learn pipelines, allowing you to use it alongside standard ML estimators.

Does this skill support deep learning embeddings?

Yes, it includes pretrained transformers for leading models such as ChemBERTa, ChemGPT, and various Graph Neural Networks (GNNs).

How does Molfeat handle large molecular datasets?

It utilizes parallel processing via the n_jobs parameter to distribute workloads across CPU cores and includes caching mechanisms to optimize repeated featurization tasks.

Molfeat Molecular Featurization

Name: Molfeat Molecular Featurization
Author: Sologa

bySologa

Data Science & ML

Transforms chemical structures into machine learning-ready numerical features using a library of over 100 featurizers and pretrained models.

About

Molfeat is a comprehensive hub for molecular featurization that unifies diverse cheminformatics tools into a single, scikit-learn compatible interface. It enables researchers to convert SMILES strings or RDKit molecules into numerical representations—ranging from traditional fingerprints like ECFP to advanced deep learning embeddings like ChemBERTa—facilitating rapid development of QSAR models, virtual screening pipelines, and chemical space visualizations. With built-in support for parallel processing and caching, it is designed to handle both small-scale research and large-scale industrial datasets efficiently.

Key Features

0 GitHub stars
Support for state-of-the-art pretrained embeddings like ChemBERTa, ChemGPT, and Graphormer
Scikit-learn compatible transformers for seamless integration into machine learning pipelines
Unified interface for 100+ featurizers including ECFP, MACCS, and Mordred descriptors
Robust error handling and validation for large-scale SMILES dataset processing
High-performance parallel processing and built-in caching for expensive computations

Use Cases

Analyzing chemical space through clustering, similarity searching, and dimensionality reduction
Developing QSAR and QSPR models for biological activity and property prediction
Executing large-scale virtual screening to rank and filter compound libraries

About

Key Features

0 GitHub stars
Support for state-of-the-art pretrained embeddings like ChemBERTa, ChemGPT, and Graphormer
Scikit-learn compatible transformers for seamless integration into machine learning pipelines
Unified interface for 100+ featurizers including ECFP, MACCS, and Mordred descriptors
Robust error handling and validation for large-scale SMILES dataset processing
High-performance parallel processing and built-in caching for expensive computations

Use Cases

Analyzing chemical space through clustering, similarity searching, and dimensionality reduction
Developing QSAR and QSPR models for biological activity and property prediction
Executing large-scale virtual screening to rank and filter compound libraries