CLIPZyme: Reaction-Conditioned Virtual Screening of Enzymes

2024/02/09 by Peter G. Mikhael, Mikhael, Peter G., Itamar Chinn +3 · 6 citations
Biochemistry, Genetics and Molecular Biology · Materials Science · #Advanced Nanomaterials in Catalysis #Amino Acid Enzymes and Metabolism #FOS: Biological sciences #Photosynthetic Processes and Mechanisms #Quantitative Methods (q-bio.QM)

paper · pdf · doi:10.48550/arxiv.2402.06748

openalex publication_date 2024/02/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Computational screening of naturally occurring proteins has the potential to identify efficient catalysts among the hundreds of millions of sequences that remain uncharacterized. Current experimental methods remain time, cost and labor intensive, limiting the number of enzymes they can reasonably screen. In this work, we propose a computational framework for in-silico enzyme screening. Through a contrastive objective, we train CLIPZyme to encode and align representations of enzyme structures and reaction pairs. With no standard computational baseline, we compare CLIPZyme to existing EC (enzyme commission) predictors applied to virtual enzyme screening and show improved performance in scenarios where limited information on the reaction is available (BEDROC₈₅ of 44.69%). Additionally, we evaluate combining EC predictors with CLIPZyme and show its generalization capacity on both unseen reactions and protein clusters.

Cited by

• Multi-Alignment Contrastive Learning for Enzyme--Reaction Retrieval
• EnzyCLIP: A Cross-Attention Dual Encoder Framework with Contrastive Learning for Predicting Enzyme Kinetic Constants
• AMix-1: A Pathway to Test-Time Scalable Protein Foundation Model
• Squidly: Enzyme Catalytic Residue Prediction Harnessing a Biology-Informed Contrastive Learning Framework
• General Protein Pretraining or Domain-Specific Designs? Benchmarking Protein Modeling on Realistic Applications
• Rethinking Benchmarks and Models for Enzyme Specificity Prediction

Related

• CARE: a Benchmark Suite for the Classification and Retrieval of Enzymes
• Multi-Alignment Contrastive Learning for Enzyme--Reaction Retrieval
• Rethinking Benchmarks and Models for Enzyme Specificity Prediction
• SCREEN: A Graph-based Contrastive Learning Tool to Infer Catalytic Residues and Assess Enzyme Mutations
• Squidly: Enzyme Catalytic Residue Prediction Harnessing a Biology-Informed Contrastive Learning Framework