Find open-source science resources

Enzymm is a tool designed to detect catalytic features (motifs) in predicted protein structures by performing geometric matching of known enzyme catalytic sites. It provides insights into enzyme function and mechanism, especially for distantly related or convergent enzymes.

QuantumPDB is a Python package and workflow that automates the generation of QM cluster models from raw protein structures, enabling high-throughput computational studies of enzyme catalysis.

DposFinder is a deep learning framework that uses transformer architecture (ESM-2) to predict phage-derived polysaccharide depolymerases and their associated host bacterial capsular serotypes. It aids in the large-scale discovery of enzymes relevant for phage therapy against antibiotic-resistant bacteria.

MechFind is a computational tool that predicts putative enzyme mechanisms using only reaction stoichiometry, enabling de novo enzyme design and significantly expanding mechanistic annotation.

EC-Bench is an open-source benchmark providing a unified framework to systematically evaluate and compare existing methods for predicting Enzyme Commission (EC) numbers from protein sequences.