Research map / Probabilistic and scientific modeling / Molecules and proteins
Towards All-Atom Foundation Models for Biomolecular Binding Affinity Prediction
Liang Shi, Zuobai Zhang, Huiyu Cai, Santiago Miret, Zhi Yang, 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 denoising pre-training (ablation) | ADiT-M · SKEMPIv2 | Pearson: 0.649 -> 0.683 | Improved |
| add denoising pre-training (ablation) | ADiT-M · SKEMPIv2 | RMSE: 1.624 -> 1.559 | Improved |
| use all-atom instead of backbone-only structures (ablation) | ADiT-M · SKEMPIv2 | Pearson: 0.658 -> 0.683 | Improved |
| use all-atom instead of backbone-only structures (ablation) | ADiT-M · SKEMPIv2 | RMSE: 1.606 -> 1.559 | Improved |
| scale model from 35M to 253M parameters (ablation) | ADiT · SKEMPIv2 | Pearson: 0.683 -> 0.691 | Improved |
| scale model from 12M to 35M parameters (ablation) | ADiT · SKEMPIv2 | Pearson: 0.660 -> 0.683 | Improved |
| scale model from 12M to 253M parameters | ADiT · Davis | MSE: 0.252 -> 0.198 | Improved |
| use ADiT instead of fine-tuned Protenix | ADiT-L · LBA-60 | RMSE: 1.781 -> 1.246 | Improved |
| use ADiT instead of fine-tuned Protenix | ADiT-L · LBA-30 | RMSE: 1.354 -> 1.308 | Improved |
| scale model size from S to L | ADiT · HER2 binders | Pearson: 0.527 -> 0.507 -> 0.567 | Mixed |
| use ADiT-L instead of GearBind + P | ADiT-L · HER2 binders | Pearson: 0.515 -> 0.567 | Improved |
| use ADiT-S instead of GEARNET-EDGE | ADiT-S · EC annotation | Fmax: 0.744 -> 0.751 (30%), 0.848 -> 0.843 (70%) | Mixed |
Closest papers by meaning
- AffinityFlow: Guided Flows for Antibody Affinity Maturation (ICML 2025 · similarity 0.89)
- Full-Atom Peptide Design with Geometric Latent Diffusion (NeurIPS 2024 · similarity 0.88)
- CausalBind: Causal Modeling and Learning for Protein-Molecule Virtual Screening (NeurIPS 2026 · similarity 0.88)
- Generative Modeling of Full-Atom Protein Conformations using Latent Diffusion on Graph Embeddings (NeurIPS 2025 · similarity 0.88)
- PPDiff: Diffusing in Hybrid Sequence-Structure Space for Protein-Protein Complex Design (ICML 2025 · similarity 0.88)
- Towards precision protein-ligand affinity prediction benchmark: A Complete and Modification-Aware DAVIS Dataset (NeurIPS 2025 · 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.
