Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Contrastive Geometric Learning Unlocks Unified Structure- and Ligand-Based Drug Design
Lisa Schneckenreiter, Sohvi Luukkonen, Lukas Friedrich, Daniel Kuhn, Günter Klambauer
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 ligand-based data to structure-based training (ablation) | ConGLUDe · LIT-PCBA | AUROC: 53.06 -> 66.25 | Improved |
| add structure-based data to ligand-based training (ablation) | ConGLUDe · LIT-PCBA | AUROC: 67.94 -> 66.25 | Mixed |
| add ligand-based data to structure-based training (ablation) | ConGLUDe · DUD-E | AUROC: 83.88 -> 82.04 | Hurt |
| add structure-based data to ligand-based training (ablation) | ConGLUDe · DUD-E | AUROC: 67.11 -> 82.04 | Improved |
| add structure-based data to ligand-based training (ablation) | ConGLUDe · Kinobeads target fishing | AUROC: 45.2 -> 65.6 | Improved |
| add ligand-based data to structure-based training (ablation) | ConGLUDe · Kinobeads target fishing | AUROC: 63.2 -> 65.6 | Improved |
| add molecule-to-binding-site loss L_m2b (ablation) | ConGLUDe · ASD pocket selection | top-1 DCC success rate: 0.10 -> 0.35 | Improved |
| add geometric loss L_geometric (ablation) | ConGLUDe · ASD pocket selection | top-1 DCC success rate: 0.01 -> 0.35 | Improved |
| add geometric loss L_geometric (ablation) | ConGLUDe · LIT-PCBA | AUROC: 64.17 -> 66.25 | Improved |
| add molecule-to-protein loss L_m2p (ablation) | ConGLUDe · LIT-PCBA | AUROC: 64.80 -> 66.25 | Improved |
| add molecule-to-protein loss L_m2p (ablation) | ConGLUDe · Kinobeads target fishing | AUROC: 65.4 -> 65.6 | No clear effect |
| add ConGLUDe ligand-conditioning over unconditioned VN-EGNN ranking | VN-EGNN · ASD pocket selection | top-1 DCC success rate: 0.28 -> 0.35 | Improved |
Closest papers by meaning
- 3DMolFormer: A Dual-channel Framework for Structure-based Drug Discovery (ICLR 2025 · similarity 0.88)
- Enhancing Ligand Validity and Affinity in Structure-Based Drug Design with Multi-Reward Optimization (ICML 2025 · similarity 0.88)
- DecoyDB: A Dataset for Graph Contrastive Learning in Protein-Ligand Binding Affinity Prediction (NeurIPS 2025 · similarity 0.87)
- Self-supervised Pocket Pretraining via Protein Fragment-Surroundings Alignment (ICLR 2024 · similarity 0.87)
- ShEPhERD: Diffusing shape, electrostatics, and pharmacophores for bioisosteric drug design (ICLR 2025 · similarity 0.87)
- TurboHopp: Accelerated Molecule Scaffold Hopping with Consistency Models (NeurIPS 2024 · similarity 0.87)
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.
