Research map / Probabilistic and scientific modeling / Molecules and proteins
Fast Proteome-Scale Protein Interaction Retrieval via Residue-Level Factorization
Jianan Zhao, Zhihao Zhan, Narendra Chaudhary, Xinyu Yuan, Zuobai Zhang, Qian Cong, Jian Zhou, Sanchit Misra, Jian Tang
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 residue-level interaction modeling (RaftPPI over RaftPPI-P) (ablation) | RaftPPI-P · 7 PPI datasets avg | AUROC: 71.90 -> 75.29 | Improved |
| add residue-level interaction modeling (RaftPPI over RaftPPI-P) (ablation) | RaftPPI-P · human proteome retrieval | Recall@20% (full): 43.83 -> 47.91 | Improved |
| add residue-level reasoning to ESM2-MLP with adaptive negative weighting loss (ablation) | ESM2-MLP-ANW-Loss · Gold | Recall@20%: -46% Recall@20% without residue-level modeling | Improved |
| add Gaussian kernel via SORF instead of linear dot product (ablation) | RaftPPI · 7 PPI datasets | AUROC and Recall@20% | Improved |
| use attention aggregation instead of averaging or CLS score (ablation) | RaftPPI · 7 PPI datasets | AUROC and Recall@20% | Improved |
| use adaptive negative weighting instead of uniform BCE (ablation) | RaftPPI · 7 PPI datasets, esp. D-SCRIPT | AUROC and Recall@20% | Improved |
| raise adversarial temperature tau from 0.125 to 4 (ablation) | RaftPPI · human and yeast | AUROC and Recall@20%: about 2-3 points AUROC, about 3 points Recall@20% | Improved |
| raise tau beyond 4 (ablation) | RaftPPI · human and yeast | AUROC and Recall@20% | Hurt |
| increase Gaussian bandwidth from 0.5 to 8.0 (ablation) | RaftPPI · human datasets | AUROC: 73.1 -> 56.5 | Hurt |
| increase attention rank from 1 to 2 (ablation) | RaftPPI · 7 datasets avg | AUROC: 75.29 -> 74.56 | Hurt |
| increase attention rank from 1 to 16 (ablation) | RaftPPI · 7 datasets avg | Recall@20%: 48.33 -> 37.91 | Hurt |
| increase attention rank from 1 to 16 (ablation) | RaftPPI · 7 datasets avg | AUROC: 75.29 -> 63.86 | Hurt |
Closest papers by meaning
- PRING: Rethinking Protein-Protein Interaction Prediction from Pairs to Graphs (NeurIPS 2025 · similarity 0.87)
- MAPE-PPI: Towards Effective and Efficient Protein-Protein Interaction Prediction via Microenvironment-Aware Protein Embedding (ICLR 2024 · similarity 0.87)
- PPDiff: Diffusing in Hybrid Sequence-Structure Space for Protein-Protein Complex Design (ICML 2025 · similarity 0.85)
- Learning to design protein-protein interactions with enhanced generalization (ICLR 2024 · similarity 0.84)
- ProtCLIP: Function-Informed Protein Multi-Modal Learning (AAAI 2025 · similarity 0.83)
- MMSite: A Multi-modal Framework for the Identification of Active Sites in Proteins (NeurIPS 2024 · similarity 0.83)
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.
