Research map / Learning theory and algorithms / Optimal transport
Fast and Interpretable Protein Substructure Alignment via Optimal Transport
Zhiyu Wang, Bingxin Zhou, Weishu Zhao, Yang Tan, Jing Wang, Pietro Lio, Liang Hong
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 |
|---|---|---|---|
| learn the cost network with supervision instead of training-free PLASMA-PF (ablation) | PLASMA (Ankh) · VenusX Motif test_extra | ROC-AUC: .93 -> .97 (ESM2) | Improved |
| train the siamese cost network (vs PLASMA-PF) (ablation) | PLASMA (ProtSSN) · VenusX Motif test_extra | LMS: .23 -> .52 | Improved |
| train the siamese cost network (vs PLASMA-PF) (ablation) | PLASMA (ProstT5) · VenusX Motif test_inter | LMS: .57 -> .91 | Improved |
| use PLASMA OT alignment vs EBA on structure-aware ProtSSN embeddings | PLASMA · VenusX Motif test_extra | ROC-AUC: .32 -> .96 | Improved |
| use PLASMA vs global structure alignment (TM-Align) | PLASMA (Ankh) · VenusX Motif test_extra | ROC-AUC: .81 -> .98 | Improved |
| use PLASMA vs TM-Align on binding site detection | PLASMA (Ankh) · VenusX Binding Site test_extra | ROC-AUC: .91 -> .99 | Improved |
| use PLASMA-PF vs PLASMA on Ankh motif detection (ablation) | PLASMA-PF · VenusX Motif test_extra | ROC-AUC: .98 vs .98 | No clear effect |
| replace EBA dynamic programming with Sinkhorn OT on GPU | PLASMA · pair inference time | time per pair: about 3 times faster than EBA | Improved |
Closest papers by meaning
- Evaluating Representation Learning on the Protein Structure Universe (ICLR 2024 · similarity 0.85)
- Protein Structure Tokenization: Benchmarking and New Recipe (ICML 2025 · similarity 0.85)
- Atom-level Protein Representation Learning Improves Protein Structure Prediction (NeurIPS 2026 · similarity 0.84)
- Pre-Training Protein Bi-level Representation Through Span Mask Strategy On 3D Protein Chains (ICML 2024 · similarity 0.84)
- AlphaFold Database Debiasing for Robust Inverse Folding (NeurIPS 2025 · similarity 0.84)
- Sidechain conditioning and modeling for full-atom protein sequence design with FAMPNN (ICML 2025 · similarity 0.84)
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.
