Research map / 3D vision / 3D generation
Zero-Shot 3D Drug Design by Sketching and Generating
Siyu Long, Yi Zhou, Xinyu Dai, Hao Zhou
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 pocket-based shape sketching instead of ligand-based sketching (ablation) | Desert · Set B | Median Vina Score (kcal/mol): -7.34 -> -9.32 | Improved |
| Use pocket-based shape sketching instead of ligand-based sketching (ablation) | Desert · Set A | Median Vina Score (kcal/mol): -8.89 -> -9.62 | Improved |
| Add post-processing (ablation) | Desert-Pocket · 55 recovered protein pockets | Median Vina Score (kcal/mol): -6.148 -> -9.410 | Improved |
| Add post-processing (ablation) | Desert-Pocket · 55 recovered protein pockets | QED: 0.614 -> 0.549 | Hurt |
| Add post-processing (ablation) | Desert-Pocket · 55 recovered protein pockets | SA: 0.612 -> 0.616 | No clear effect |
| Add post-processing (ablation) | Desert-Pocket · 55 recovered protein pockets | Diversity: 0.926 -> 0.908 | Hurt |
| Increase the number of molecules sampled before post-processing (ablation) | Desert-Pocket · Set B | performance | Improved |
| Increase the pre-trained model size (ablation) | Shape2Mol · ZINC extra test set | model quality | Improved |
| Increase pre-training dataset size beyond a moderate size (ablation) | Shape2Mol · ZINC extra test set | model quality | No clear effect |
| Use shape-guided generation instead of database screening | Desert-Pocket · Set B | Median Vina Score (kcal/mol): -7.95 -> -9.32 | Improved |
| Use shape-guided generation instead of database screening | Desert-Pocket · Set A | Median Vina Score (kcal/mol): -8.66 -> -9.62 | Improved |
| Use pocket-based sketching rather than only the reference ligand region (ablation) | Desert · Set B | Success rate (%): 18.0 -> 61.0 | Improved |
Closest papers by meaning
- Controllable 3D Molecular Generation for Structure-Based Drug Design Through Bayesian Flow Networks and Gradient Integration (NeurIPS 2025 · similarity 0.84)
- ShEPhERD: Diffusing shape, electrostatics, and pharmacophores for bioisosteric drug design (ICLR 2025 · similarity 0.84)
- DrugHash: Hashing Based Contrastive Learning for Virtual Screening (AAAI 2025 · similarity 0.83)
- CBGBench: Fill in the Blank of Protein-Molecule Complex Binding Graph (ICLR 2025 · similarity 0.83)
- Navigating the Design Space of Equivariant Diffusion-Based Generative Models for De Novo 3D Molecule Generation (ICLR 2024 · similarity 0.82)
- Uncertainty-Aware Multi-Objective Reinforcement Learning-Guided Diffusion Models for 3D De Novo Molecular Design (NeurIPS 2025 · 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.
