Research map / Vision-language and multimodal / Human and embodied modeling
Multimodal Protein Language Models for Enzyme Kinetic Parameters: From Substrate Recognition to Conformational Adaptation
Fei Wang, Xinye Zheng, Kun Li, Yanyan Wei, Yuxin Liu, Ganpeng Hu, Tong Bao, Jingwen Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add MRCA substrate cross-attention to the PLM (ablation) | ESM2-3B PLM baseline · Ki prediction | R2: 0.49 -> 0.51 | Improved |
| add G-MoE geometry-aware experts after MRCA (ablation) | ERBA · Ki prediction | R2: 0.51 -> 0.54 | Improved |
| add ESDA distribution alignment after G-MoE (ablation) | ERBA · Ki prediction | R2: 0.54 -> 0.61 | Improved |
| use MRCA instead of concat and self-attention fusion (ablation) | ERBA · Ki prediction | R2: 0.47 -> 0.61 | Improved |
| use G-MoE instead of standard MoE (ablation) | ERBA · Ki prediction | R2: 0.50 -> 0.61 | Improved |
| use ESDA objective instead of plain L2 loss (ablation) | ERBA · Ki prediction | R2: 0.48 -> 0.61 | Improved |
| use MRCA instead of concat and self-attention fusion (ablation) | ERBA · Ki prediction | RMSE: 1.44 -> 1.26 | Improved |
| fuse enzyme, then substrate, then structure instead of enzyme, structure, substrate (ablation) | ERBA · Ki prediction | R2: 37.8% R2 improvement, 20% RMSE reduction | Improved |
| attach ERBA to ESM2-3B backbone | ESM2-3B · kcat prediction | R2: 0.41 -> 0.54 | Improved |
| attach ERBA to ESM2-3B backbone | ESM2-3B · Km prediction | R2: 0.43 -> 0.61 | Improved |
| attach ERBA to ESM2-8M backbone | ESM2-8M · kcat prediction | R2: 0.19 -> 0.27 | Improved |
| attach ERBA to Ankh3-5.7B backbone | Ankh3-5.7B · Ki prediction | R2: 0.46 -> 0.57 | Improved |
Closest papers by meaning
- Generative Enzyme Design Guided by Functionally Important Sites and Small-Molecule Substrates (ICML 2024 · similarity 0.87)
- EnzyControl: Adding Functional and Substrate-Specific Control for Enzyme Backbone Generation (NeurIPS 2025 · similarity 0.86)
- ReactZyme: A Benchmark for Enzyme-Reaction Prediction (NeurIPS 2024 · similarity 0.86)
- Retrieval Augmented Zero-Shot Enzyme Generation for Specified Substrate (ICML 2025 · similarity 0.86)
- Simultaneous Modeling of Protein Conformation and Dynamics via Autoregression (NeurIPS 2025 · similarity 0.86)
- Unlocking hidden biomolecular conformational landscapes in diffusion models at inference time (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
