Research map / Image generation / Autoregressive and latent generation
AbDiffuser: full-atom generation of in-vitro functioning antibodies
Karolis Martinkus, Jan Ludwiczak, WEI-CHING LIANG, Julien Lafrance-Vanasse, Isidro Hotzel, Arvind Rajpal, Yan Wu, Kyunghyun Cho, Richard Bonneau, Vladimir Gligorijevic, Andreas Loukas
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 AHo numbering and position-specific residue-frequency priors (ablation) | antibody generation models · Paired OAS | W1(Nat.): 0.2837 -> 0.1979 | Improved |
| use learned conditional atom covariance (ablation) | AbDiffuser · Paired OAS | RMSD: 0.6302 -> 0.5230 | Improved |
| remove the physics-based projection layer (ablation) | AbDiffuser · Paired OAS | RMSD: 0.5230 -> 11.1431 | Hurt |
| generate side chains alongside the backbone (ablation) | AbDiffuser · Paired OAS | W1(Nat.): 0.1979 -> 0.0916 | Improved |
| generate side chains alongside the backbone (ablation) | AbDiffuser · HER2 binder generation | p_bind: 0.5761 -> 0.6848 | Improved |
| generate side chains alongside the backbone (ablation) | AbDiffuser · HER2 binder generation | W1(Sta.): 0.0028 -> 0.0062 | Hurt |
| reduce sampling temperature to τ = 0.75 (ablation) | AbDiffuser · HER2 binder generation | p_bind: 0.5761 -> 0.6382 | Improved |
| reduce sampling temperature to τ = 0.01 (ablation) | AbDiffuser · HER2 binder generation | Uniq.: 100% -> 99.7% | Hurt |
| reduce sampling temperature to τ = 0.01 (ablation) | AbDiffuser (side chains) · HER2 binder generation | Uniq.: 100% -> 91.4% | Hurt |
| filter generated designs by in-silico metrics and classifier confidence | AbDiffuser HER2 designs · HER2 in-vitro designs | binding rate: 22. 2% -> 57. 1% | Improved |
Closest papers by meaning
- Antibody Design Using a Score-based Diffusion Model Guided by Evolutionary, Physical and Geometric Constraints (ICML 2024 · similarity 0.90)
- Retrieval Augmented Diffusion Model for Structure-informed Antibody Design and Optimization (ICLR 2025 · similarity 0.89)
- Antigen-Specific Antibody Design via Direct Energy-based Preference Optimization (NeurIPS 2024 · similarity 0.87)
- AffinityFlow: Guided Flows for Antibody Affinity Maturation (ICML 2025 · similarity 0.87)
- Protein Discovery with Discrete Walk-Jump Sampling (ICLR 2024 · similarity 0.87)
- Cross-Gate MLP with Protein Complex Invariant Embedding Is a One-Shot Antibody Designer (AAAI 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.
