Research map / Label-efficient and robust learning / Recognition and long-tail
LipoPU: Pocket-level Prediction of Lipid-Protein Interactions via Positive-Unlabeled Learning
Yuxing Wang, Wenyi Zhang, Yilong Zou, Jing Huang
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 |
|---|---|---|---|
| use ranking-based PU objective instead of weighted BCE | LipoPU · Easy test set | Binary PR-AUC: 0.630 -> 0.839 | Improved |
| use ranking-based PU objective instead of weighted BCE | LipoPU · Easy test set | Micro PR-AUC: 0.453 -> 0.766 | Improved |
| use ranking-based PU objective instead of weighted BCE | LipoPU · Hard test set | Binary PR-AUC: 0.557 -> 0.650 | Improved |
| use ranking-based PU objective instead of weighted BCE | LipoPU · Hard test set | Micro PR-AUC: 0.195 -> 0.280 | Improved |
| add positive-guaranteed batch sampler to BCE | BCE baseline · Easy test set | Binary PR-AUC: 0.630 -> 0.651 | No clear effect |
| use LipoPU instead of SLiPP | pocket-level lipid binding · Easy-fpocket | Binary PR-AUC: 0.239 -> 0.792 | Improved |
| use LipoPU instead of SLiPP | pocket-level lipid binding · Hard-fpocket | Binary ROC-AUC: 0.726 -> 0.881 | Improved |
| add positive-guaranteed batch sampler (ablation) | LipoPU · Easy test set | Micro PR-AUC: 0.696 -> 0.766 | Improved |
| add positive-guaranteed batch sampler (ablation) | LipoPU · Hard test set | Micro PR-AUC: 0.250 -> 0.280 | Improved |
| use ranking-based PU loss instead of standard PU loss (ablation) | LipoPU · Easy test set | Binary PR-AUC: 0.449 -> 0.839 | Improved |
| use ranking-based PU loss instead of standard PU loss (ablation) | LipoPU · Hard test set | Micro PR-AUC: 0.208 -> 0.280 | Improved |
| use attention MIL pooling instead of mean pooling (ablation) | LipoPU · Easy test set | Micro PR-AUC: 0.731 -> 0.766 | Improved |
Closest papers by meaning
- ProtCLIP: Function-Informed Protein Multi-Modal Learning (AAAI 2025 · similarity 0.83)
- Self-supervised Pocket Pretraining via Protein Fragment-Surroundings Alignment (ICLR 2024 · similarity 0.82)
- Generalized Protein Pocket Generation with Prior-Informed Flow Matching (NeurIPS 2024 · similarity 0.81)
- Knowledge Enhanced Representation Learning for Drug Discovery (AAAI 2024 · similarity 0.81)
- UniSite: The First Cross-Structure Dataset and Learning Framework for End-to-End Ligand Binding Site Detection (NeurIPS 2025 · similarity 0.81)
- PSC-CPI: Multi-Scale Protein Sequence-Structure Contrasting for Efficient and Generalizable Compound-Protein Interaction Prediction (AAAI 2024 · similarity 0.81)
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.
