Каталог generative tools
Фильтры работают локально, без сервера. Canonical description каждого инструмента — на его странице, без дублей.
| Инструмент | Что генерирует | Класс | Публикация | Evidence | Repo / service |
|---|---|---|---|---|---|
| AbDiffuser | Full-atom antibody sequence + structure | joint-sequence-structure | Preprint | 3 prospective binding + expression validation | NR |
| AlphaProteo | Binder sequences/structures | target-conditioned-binder | Preprint Technical report | 3–4 prospective binding/functional validation | https://github.com/google-deepmind/ap_trusted_testers |
| AntiFold | Fv/VHH sequences, per-residue probabilities, sampled masked variants | inverse-folding | Peer-reviewed | 2 retrospective antibody evaluation | https://github.com/oxpig/AntiFold |
| BindCraft | Binder sequence and designed target–binder complex | target-conditioned-binder | Peer-reviewed | 3–4 prospective binding and functional validation | https://github.com/martinpacesa/BindCraft |
| BoltzGen | Binder sequence–structure and target complex | joint-sequence-structure; target-conditioned-binder | Preprint | 3–4 prospective binding/functional validation | https://github.com/HannesStark/boltzgen |
| Chroma | Backbone/all-atom structure and sequence | backbone-structure-generator; joint-sequence-structure | Peer-reviewed | 3–5 early prospective biophysical/structural validation | https://github.com/generatebio/chroma |
| DiffAb | CDR sequence + backbone/geometry; interface proposal | joint-sequence-structure; target-conditioned-binder | NeurIPS 2022 main track (A*) | 1–2 in silico / retrospective | https://github.com/luost26/diffab |
| dyMEAN | Full-atom antibody sequence/structure and docked complex | joint-sequence-structure; target-conditioned-binder | ICML 2023 main track (A*) | 1–2 in silico / retrospective | https://github.com/thunlp-mt/dymean |
| Efficient evolution with ESM-1v | Ranked single-amino-acid substitutions | mutation-optimizer | Peer-reviewed | 4 prospective functional validation | https://github.com/brianhie/efficient-evolution |
| ESM-IF1 | Sequence and sequence likelihood conditioned on backbone | inverse-folding | ICML 2022 main track (A*) | 2 retrospective in primary paper; 4 when embedded in later structural-evolution workflow | https://github.com/facebookresearch/esm |
| ESM3 | Sequence, structure tokens, function annotations; can co-generate modalities | sequence-generator; joint-sequence-structure | Peer-reviewed | 4 prospective functional validation | https://github.com/Biohub/esm |
| EvoDiff | Amino-acid sequence | sequence-generator | Preprint | 1–2 in silico / retrospective | https://github.com/microsoft/evodiff |
| EVOLVEpro | Ranked single or multi-mutant variants for next round | mutation-optimizer | Peer-reviewed | 4 prospective functional validation | https://github.com/mat10d/EvolvePro |
| FlowDesign | CDR sequence and structure | joint-sequence-structure | Peer-reviewed | 3–4 prospective binding/neutralization validation | NR |
| FrameDiff | Protein backbone | backbone-structure-generator | ICML 2023 main track (A*) | 1 in silico generation/evaluation | https://github.com/jasonkyuyim/se3_diffusion |
| HuAbDiffusion | Whole humanized V-region sequences | humanization | Peer-reviewed | 2–3 retrospective + prospective binding claim | https://www.genscript.com/yabXnization-heterogenization-tool.html |
| HuDiff | Humanized variable-domain sequence around supplied CDRs | humanization | Preprint | 3 prospective binding validation | https://github.com/TencentAI4S/HuDiff |
| Humatch | Humanized heavy/light variable-domain sequences and mutation set | humanization | Peer-reviewed | 2 retrospective evaluation on 177 humanized antibodies | https://github.com/oxpig/Humatch |
| IgLM | Full heavy/light-chain sequence or infilled span/CDR | sequence-generator | Peer-reviewed | 2 retrospective antibody evaluation | https://github.com/Graylab/IgLM |
| LigandMPNN | Sequence + side-chain conformations | inverse-folding | Peer-reviewed | 3–4 prospective experimental validation | https://github.com/dauparas/LigandMPNN |
| Makowski et al. affinity–specificity co-optimization | Combinatorial CDR mutation sets predicted to trade affinity against nonspecific binding | mutation-optimizer | Peer-reviewed | 3–5 prospective binding/specificity validation | NR |
| PALM-H3 | CDR-H3 sequence; grafted antibody candidates | sequence-generator; target-conditioned-binder | Peer-reviewed | 3–4 prospective binding and functional validation | https://github.com/TencentAILabHealthcare/PALM |
| ProGen / ProGen2 | Amino-acid sequence | sequence-generator | Peer-reviewed | 4 prospective functional validation | https://github.com/salesforce/progen |
| ProteinGenerator | Coupled sequence–structure proposals | joint-sequence-structure | Peer-reviewed | 3–4 prospective functional validation | https://github.com/RosettaCommons/protein_generator |
| ProteinMPNN | One or many sequences compatible with input backbone | inverse-folding | Peer-reviewed | 3–4 prospective structural and functional validation | https://github.com/dauparas/ProteinMPNN |
| RFantibody | CDR backbones, CDR sequences, antibody–target complex proposals | backbone-structure-generator; target-conditioned-binder | Peer-reviewed | 3–4 prospective binding and functional validation | https://github.com/RosettaCommons/RFantibody |
| RFdiffusion All-Atom (RFdiffusionAA) | Protein backbone positioned around non-protein atoms; sequence typically downstream | backbone-structure-generator | Peer-reviewed | 3–4 prospective structural/functional validation | https://github.com/baker-laboratory/rf_diffusion_all_atom |
| RFdiffusion | Protein backbone / complex geometry; sequence is downstream | backbone-structure-generator; target-conditioned-binder | Peer-reviewed | 3–4 prospective binding/functional/structural validation | https://github.com/RosettaCommons/RFdiffusion |
| Structural evolution (ESM-IF1 complex) | Ranked sequence mutations preserving complex structure | mutation-optimizer | Peer-reviewed | 4 prospective functional validation | https://github.com/varun-shanker/structural-evolution |
| tFold System | CDR sequences, candidate antibody structures and antibody–antigen complexes | target-conditioned-binder | Peer-reviewed | 3–5 prospective binding, epitope and developability evidence | https://github.com/TencentAI4S/tfold |