Research map / Vision-language and multimodal / Cross-modal learning
Multimodal Mixture-of-Experts with Retrieval Augmentation for Protein Active Site Identification
Jiayang Wu, Jiale Zhou, Rubo Wang, Xingyi Zhang, Xun Lin, Tianxu Lv, Leong Hou U, Yefeng Zheng
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 reliability-aware multimodal fusion (RMF) (ablation) | MERA · ProTAD-Gen | AUPRC: 0.83 -> 0.90 | Improved |
| add RMF (ablation) | MERA · ProTAD-Gen | Fmax: 0.70 -> 0.88 | Improved |
| add text modality (ablation) | MERA · ProTAD-Gen | Fmax: 0.83 -> 0.88 | Improved |
| add RAG modality (ablation) | MERA · ProTAD-Gen | Fmax: 0.79 -> 0.88 | Improved |
| use residue-level MeRAG gating instead of direct expert combination (ablation) | MERA · ProTAD-Gen | Fmax: 0.76 -> 0.88 | Improved |
| add sequence expert (ablation) | MERA · ProTAD-Gen | Fmax: 0.83 -> 0.88 | Improved |
| add chain expert (ablation) | MERA · ProTAD-Gen | Fmax: 0.84 -> 0.88 | Improved |
| add active-site expert (ablation) | MERA · ProTAD-Gen | Fmax: 0.85 -> 0.88 | Improved |
| add peptide expert (4 experts vs 3) (ablation) | MERA · TS125 | Fmax: 0.37 -> 0.40 | Improved |
| add retrieval and text modalities to sequence-only ESM-1b features (ablation) | MERA · ProTAD-Gen | AUPRC: 0.85 -> 0.90 | Improved |
| MERA vs strongest baseline MMSite | MERA · ProTAD-Gen | Fmax: 0.81 -> 0.88 | Improved |
| MERA vs IIDL-PepPI on peptide binding | MERA · TS125 | AUROC: 0.84 -> 0.85 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MMSite — MMSite: A Multi-modal Framework for the Identification of Active Sites in Proteins (NeurIPS 2024) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MMSite: A Multi-modal Framework for the Identification of Active Sites in Proteins (NeurIPS 2024 · similarity 0.87)
- Knowledge Enhanced Representation Learning for Drug Discovery (AAAI 2024 · similarity 0.84)
- ProtCLIP: Function-Informed Protein Multi-Modal Learning (AAAI 2025 · similarity 0.84)
- Latent Retrieval Augmented Generation of Cross-Domain Protein Binders (NeurIPS 2025 · similarity 0.84)
- CausalBind: Causal Modeling and Learning for Protein-Molecule Virtual Screening (NeurIPS 2026 · similarity 0.83)
- CrossBind: Collaborative Cross-Modal Identification of Protein Nucleic-Acid-Binding Residues (AAAI 2024 · similarity 0.82)
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.
