Does it support deep learning models?

Yes, Molfeat includes pre-trained transformers for popular deep learning models like ChemBERTa, ChemGPT, and various Graph Neural Networks (GNNs).

What is Molfeat used for?

Molfeat is used to convert chemical structures (SMILES) into numerical vectors (features) that machine learning models can understand, supporting over 100 different featurization methods.

Is Molfeat optimized for large datasets?

Yes, it features built-in support for parallel processing (n_jobs) and caching to efficiently handle large-scale chemical databases.

Can I use Molfeat with scikit-learn?

Absolutely. Molfeat's transformers are scikit-learn compatible, allowing you to include featurization steps directly within your standard ML pipelines.

How does it handle invalid SMILES strings?

Molfeat allows for graceful error handling using the 'ignore_errors' flag, which lets the pipeline continue processing while logging details about failed molecules.

Molfeat Molecular Featurization

Name: Molfeat Molecular Featurization
Author: x-cmd

byx-cmd

•

データサイエンスとML

Simplifies molecular featurization by providing access to over 100 pre-trained embeddings and hand-crafted featurizers for machine learning.

Molfeat is a comprehensive Python library designed to bridge the gap between chemical structures and machine learning models. It unifies a vast array of featurizers—ranging from traditional fingerprints like ECFP and MACCS to state-of-the-art deep learning embeddings like ChemBERTa and Graphormer—into a single, scikit-learn compatible interface. Whether you are performing virtual screening, building QSAR models, or analyzing chemical space, this skill provides the necessary tools to convert SMILES strings or RDKit molecules into high-quality numerical representations with parallel processing and built-in caching.

主な機能

01Scikit-learn compatible transformers for seamless ML pipeline integration

02Unified interface for 100+ molecular featurizers and pre-trained embeddings

03Supports diverse representations including fingerprints, 2D descriptors, and GNNs

04High-performance batch processing with automatic parallelization and caching

058 GitHub stars

06State management for saving and loading exact featurizer configurations

ユースケース

01Building QSAR/QSPR models for drug discovery and molecular property prediction

02Visualizing and clustering chemical space for exploratory data analysis

03Conducting large-scale virtual screening and similarity searches in chemical libraries

主な機能

01Scikit-learn compatible transformers for seamless ML pipeline integration

02Unified interface for 100+ molecular featurizers and pre-trained embeddings

03Supports diverse representations including fingerprints, 2D descriptors, and GNNs

04High-performance batch processing with automatic parallelization and caching

058 GitHub stars

06State management for saving and loading exact featurizer configurations

ユースケース

01Building QSAR/QSPR models for drug discovery and molecular property prediction

02Visualizing and clustering chemical space for exploratory data analysis

03Conducting large-scale virtual screening and similarity searches in chemical libraries