EvoZymePro-Cat: A Protein–Ligand-Aware Deep Learning Framework for Predicting Mutation Effects in Enzyme Function
Bibliographic record
Abstract
Enzymes are biological catalysts that speed up chemical reactions in an eco-friendly way. Precise enzyme design is hindered by vast sequence space and intricate sequence–structure–function interdependencies. To address these challenges, we developed EvoZymePro-Cat (EZPro-Cat), a deep learning platform for enzyme mutant screening. Conventional methods for predicting absolute mutant activities suffer from systematic errors and limited generalizability. Our pairwise comparison framework directly models relative activity superiority between variants, eliminating dependence on absolute value predictions. The framework integrates full sequence and local structure semantics of protein and ligand information using bilinear attention mechanisms. Protein sequences are encoded using the ESM1b transformer model. Ligands are represented through MolT5 embeddings and MACCS molecular fingerprints. The adaptability of protein residues to their microenvironments is captured by integrating structural features and site-specific evolutionary characteristics. Bilinear attention mechanisms capture long-range intermolecular interactions during catalysis by bidirectional projection and weighted fusion of protein–ligand features. Compared to existing methods, our model exhibits superior performance in identifying improved enzyme mutants through comparative prediction of mutation effects on activity, such as K m and k cat . For deep mutation scanning data sets, a few-shot learning strategy combined with the EZPro-Cat framework boosts prediction precision (AUC 0.908). By using integrated multimodal representations, EZPro-Cat offers a mechanistic and practical solution for functional profiling of intraprotein variants, driving paradigm shifts in highly efficient enzyme discovery and directed evolution.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".