Главная / Инструменты / RFdiffusion All-Atom (RFdiffusionAA)

RFdiffusion All-Atom (RFdiffusionAA)

RFdiffusionAA RFdiffusion All-Atom

Peer-reviewed Prospective flag

Summary card

Поле	Значение
Год / version	2024 / NR
Last verified	2026-08-13
Что генерирует	Protein backbone positioned around non-protein atoms; sequence typically downstream
Форматы	Protein–small molecule / nucleic acid / metal systems
De novo или optimization	De novo structure
DOI	10.1126/science.adl2528
Paper	https://doi.org/10.1126/science.adl2528
GitHub / repo	https://github.com/baker-laboratory/rf_diffusion_all_atom
Service / docs	NR
Лицензия	Repository license marked Other; exact commercial terms require review
Доступность / commercial-use	Code + weights/container

Задача и generated object

Used in these tasks: De novo генерация backbone, Дизайн функционального сайта

Part of these workflows: De novo backbone → inverse folding

Поле	Значение
Required input	Ligand/cofactor/metal/nucleic-acid context; optional motif
Conditioning	Atomic motifs and heterogeneous biomolecular context
Fixed	NR
Editable	NR
Generated	Protein backbone positioned around non-protein atoms; sequence typically downstream
Single mutations	No
Combinatorial mutations	Via downstream sequence design
Multi-objective	No general multi-objective loop

Architecture and operating principle

Author-reported (matrix)

Поле	Значение
Base architecture	NR
Model type	3 Backbone/structure generator
Representation	NR
Generation unit	Protein backbone positioned around non-protein atoms; sequence typically downstream
Conditioning	Atomic motifs and heterogeneous biomolecular context
Output	Protein backbone positioned around non-protein atoms; sequence typically downstream
Sampling	NR
Internal scorer	No universal ranker
Structure generation	Uses all-atom learned representation; generation remains distinct from validation

Input representation

Поле	Значение
Node features	NR
Edge / pair features	NR

Encoder и decoder

Поле	Значение
Encoder	NR
Decoder	NR

Latent or generative state

Conditioning mechanism

Atomic motifs and heterogeneous biomolecular context

Generation mechanism

All-atom RoseTTAFold-derived diffusion over biomolecular context

Sampling: NR

Training objectives

Losses: NR

Mathematical details: NR

Integrated evaluation

No universal ranker

Architectural limitations

Legacy environment/container.
generates structure before sequence.
ligand pose and chemistry assumptions remain critical.

Uses these architecture components: Coordinate diffusion, SE(3)-equivariant networks

Generator versus evaluation

Component	Role	Architecture class	Integrated or external	Required or optional	Influences generation or post hoc	Runtime included or excluded
RFdiffusionAA	primary_generator	All-atom RoseTTAFold-derived diffusion over biomolecular context	integrated	required	generation	included
No universal ranker	integrated_scorer	NR	integrated	optional	post_hoc	unknown
Uses all-atom learned representation; generation remains distinct from validation	integrated_structure_predictor	NR	integrated	optional	post_hoc	included
LigandMPNN/ProteinMPNN, folding, pose/energy/experimental validation	mandatory_external_dependency	NR	external	required	post_hoc	excluded
Experimental binding/function assays	experimental_validation	NR	external	required	post_hoc	excluded

Training data и biases

Поле	Значение
Training data	NR
Preprocessing	NR
Negative examples	NR
Fine-tuning	NR
Label requirements	NR
Training hardware	NR
Training time	NR

Inference hardware и runtime

Поле	Значение
Candidate count	NR
Batch size	NR
Generation steps	NR
Recycles	NR
Size	NR
Reported runtime	NR
Runtime boundary	NR
CPU	NR
GPU	NR
VRAM	NR
RAM	NR
Disk	NR
Databases	NR

Experimental evidence

Поле	Значение
Evidence class	3–4 prospective structural/functional validation
Retrospective	NR
Prospective binding	General biomolecular design evidence
Prospective function	General biomolecular design evidence
Developability	NR
Tested design count	Multiple small-molecule-centered design campaigns; aggregate count: NR
Success criterion	General biomolecular design evidence
Independent validation	NR
Summary	General biomolecular design evidence

Limitations and failure modes

Legacy environment/container.
generates structure before sequence.
ligand pose and chemistry assumptions remain critical.

Deployment: Limited pilot; deployment friction high

Reproducibility, license, commercial use

Поле	Значение
Repo	https://github.com/baker-laboratory/rf_diffusion_all_atom
License	Repository license marked Other; exact commercial terms require review
Commercial-use status	Code + weights/container

Alternative tools: NR

Complementary: NR

Related: NR

Not directly comparable with

AntiFold, Efficient evolution with ESM-1v, ESM-IF1, EvoDiff, LigandMPNN, ProGen / ProGen2, ProteinMPNN, Structural evolution (ESM-IF1 complex)

Сравнение допустимо только внутри совместимых категорий. Runtime и quality AntiFold vs RFantibody нельзя читать как одну шкалу: разные starting points и generated objects.

Depends on these external modules

Primary paper и полный список источников

Generalized biomolecular modeling and design with RoseTTAFold All-Atom

Source matrix: landscape xlsx, last verified 2026-08-13.

Детальный source-grounded разбор

1. Что именно делает инструмент

[MATRIX]Вход: Ligand/cofactor/metal/nucleic-acid context; optional motif

[MATRIX]Generated object: Protein backbone positioned around non-protein atoms; sequence typically downstream

[MATRIX]Фиксируется: NR. Изменяется: NR.

[MATRIX]De novo или parent-based: De novo structure.

2. Архитектура в одном абзаце

[MATRIX]All-atom RoseTTAFold-derived diffusion over biomolecular context

4. Математическая модель

NRТочные уравнения, losses и sampling solver будут внесены после чтения paper/code. Не подставляйте родовой UCB/transformer/diffusion, если он не подтверждён для этого инструмента.

Поля NR: conditional probability, update equations, sampling algorithm

5. Физическое устройство состояния

[MATRIX]Seed generated state / output: Protein backbone positioned around non-protein atoms; sequence typically downstream. Structure generation: Uses all-atom learned representation; generation remains distinct from validation.

6. Физико-химическая интерпретация

[INTERPRETATION]Координаты, frames, PLM embeddings или residue tokens не являются расчётом ΔG, электростатики или растворителя, пока это не сделано внешней energy function. Likelihood / pLDDT / docking score ≠ экспериментальный binding.

7. Training

[MATRIX]Данные: NR. Fine-tuning: NR. Negative examples: NR.

8. Inference

[MATRIX]Sampling: NR. Generation steps: NR. Runtime boundary: NR.

10. Почему, по мнению авторов, метод эффективен

[MATRIX]Author-reported summary (матрица): General biomolecular design evidence

NRCausal support не классифицирован, пока не прочитаны ablation и experimental sections.

12. Ablation analysis

NRAblation в этом выпуске не разобрана.

13. What the model does not represent

Явный растворитель и ионная сила — обычно отсутствуют в генераторе.
Протонирование, гликозилирование, кофакторы — обычно отсутствуют, если не оговорено.
Конформационная энтропия и долгая динамика — не моделируются одним forward pass.
Экспрессия, агрегация, иммуногенность — не равны likelihood генератора.
Истинная binding free energy — не равна internal score.

14. Evidence and limitations

[MATRIX]Evidence class (матрица): 3–4 prospective structural/functional validation. Prospective binding: General biomolecular design evidence. Function: General biomolecular design evidence. Developability: NR.

Legacy environment/container.
generates structure before sequence.
ligand pose and chemistry assumptions remain critical.

15. Reproducibility

NRРепозиторий: https://github.com/baker-laboratory/rf_diffusion_all_atom. Лицензия: Repository license marked Other; exact commercial terms require review. Commit: NR, пока страница не имеет отдельного analysis YAML.

Модульная схема

Модуль	Класс модели	Вход	Выход	Обучается или frozen	Generator или evaluator
RFdiffusionAA	All-atom RoseTTAFold-derived diffusion over biomolecular context	NR	NR	NR	primary_generator
No universal ranker	NR	NR	NR	NR	integrated_scorer
Uses all-atom learned representation; generation remains distinct from validation	NR	NR	NR	NR	integrated_structure_predictor
LigandMPNN/ProteinMPNN, folding, pose/energy/experimental validation	NR	NR	NR	NR	mandatory_external_dependency
Experimental binding/function assays	NR	NR	NR	NR	experimental_validation

Generator versus evaluator

Компонент	Роль	Integrated или external	Required или optional	Влияет на generation или только ранжирует после
RFdiffusionAA	primary_generator	integrated	required	generation
No universal ranker	integrated_scorer	integrated	optional	post_hoc
Uses all-atom learned representation; generation remains distinct from validation	integrated_structure_predictor	integrated	optional	post_hoc
LigandMPNN/ProteinMPNN, folding, pose/energy/experimental validation	mandatory_external_dependency	external	required	post_hoc
Experimental binding/function assays	experimental_validation	external	required	post_hoc

Claim–evidence matrix

Утверждение	Предполагаемый механизм	Тип evidence	Датасет / эксперимент	Comparator	Метрика / результат	Источник	Оговорка
General biomolecular design evidence	All-atom RoseTTAFold-derived diffusion over biomolecular context	5. prospective functional validation	Multiple small-molecule-centered design campaigns; aggregate count: NR	NR	General biomolecular design evidence	[MATRIX] landscape matrix 2026-08-13	Не читать как полный разбор primary paper.