Chroma

Peer-reviewed Prospective flag

Summary card

Поле	Значение
Год / version	2023 / NR
Last verified	2026-08-13
Что генерирует	Backbone/all-atom structure and sequence
Форматы	Monomers and large protein complexes
De novo или optimization	Primarily de novo
DOI	10.1038/s41586-023-06728-8
Paper	https://doi.org/10.1038/s41586-023-06728-8
GitHub / repo	https://github.com/generatebio/chroma
Service / docs	NR
Лицензия	Apache-2.0 code; model download requires API key/terms
Доступность / commercial-use	Code + gated model access

Задача и generated object

Used in these tasks: De novo генерация backbone, Joint sequence–structure generation, Генерация полной 3D structure

Part of these workflows: Joint / multimodal sequence–structure generation

Поле	Значение
Required input	None or programmable conditioners
Conditioning	Substructure, symmetry, shape, class/annotation and custom differentiable conditioners
Fixed	NR
Editable	NR
Generated	Backbone/all-atom structure and sequence
Single mutations	Possible through constraints, not mutation-first
Combinatorial mutations	Yes
Multi-objective	Multiple conditioners, not an experimentally validated Pareto optimizer

Architecture and operating principle

Author-reported (matrix)

Поле	Значение
Base architecture	NR
Model type	3 + 4 Backbone and joint sequence–structure generator
Representation	NR
Generation unit	Backbone/all-atom structure and sequence
Conditioning	Substructure, symmetry, shape, class/annotation and custom differentiable conditioners
Output	Backbone/all-atom structure and sequence
Sampling	NR
Internal scorer	Conditioner energies / model likelihood
Structure generation	Internal structural generation; not an independent validator

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

Substructure, symmetry, shape, class/annotation and custom differentiable conditioners

Generation mechanism

Backbone diffusion + equivariant graph networks + conditional random field sequence design

Sampling: NR

Training objectives

Losses: NR

Mathematical details: NR

Integrated evaluation

Conditioner energies / model likelihood

Architectural limitations

Weights access is gated.
conditioner scores are not measured function.
no antibody-specific proof.

Uses these architecture components: Coordinate diffusion, Graph neural networks, Joint sequence–structure generation

Generator versus evaluation

Component	Role	Architecture class	Integrated or external	Required or optional	Influences generation or post hoc	Runtime included or excluded
Chroma structure diffusion + sequence module	primary_generator	Backbone diffusion + equivariant graph networks + conditional random field sequence design	integrated	required	generation	included
Conditioner energies / model likelihood	integrated_scorer	NR	integrated	optional	post_hoc	unknown
Internal structural generation; not an independent validator	integrated_structure_predictor	NR	integrated	optional	post_hoc	included
Independent folding/complex and 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	3–5 early prospective biophysical/structural validation
Retrospective	NR
Prospective binding	General protein evidence
Prospective function	General protein evidence
Developability	General protein evidence
Tested design count	Early biophysical and crystallographic validation; exact count: NR
Success criterion	General protein evidence
Independent validation	NR
Summary	General protein evidence

Limitations and failure modes

Weights access is gated.
conditioner scores are not measured function.
no antibody-specific proof.

Deployment: Limited pilot; legal/access review

Reproducibility, license, commercial use

Поле	Значение
Repo	https://github.com/generatebio/chroma
License	Apache-2.0 code; model download requires API key/terms
Commercial-use status	Code + gated model access

Alternative tools: NR

Complementary: NR

Related: NR

Not directly comparable with

Efficient evolution with ESM-1v, 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

Illuminating protein space with a programmable generative model

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

Детальный source-grounded разбор

1. Что именно делает инструмент

[MATRIX]Вход: None or programmable conditioners

[MATRIX]Generated object: Backbone/all-atom structure and sequence

[MATRIX]Фиксируется: NR. Изменяется: NR.

[MATRIX]De novo или parent-based: Primarily de novo.

2. Архитектура в одном абзаце

[MATRIX]Backbone diffusion + equivariant graph networks + conditional random field sequence design

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: Backbone/all-atom structure and sequence. Structure generation: Internal structural generation; not an independent validator.

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 protein 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): 3–5 early prospective biophysical/structural validation. Prospective binding: General protein evidence. Function: General protein evidence. Developability: General protein evidence.

Weights access is gated.
conditioner scores are not measured function.
no antibody-specific proof.

15. Reproducibility

NRРепозиторий: https://github.com/generatebio/chroma. Лицензия: Apache-2.0 code; model download requires API key/terms. Commit: NR, пока страница не имеет отдельного analysis YAML.

Модульная схема

Модуль	Класс модели	Вход	Выход	Обучается или frozen	Generator или evaluator
Chroma structure diffusion + sequence module	Backbone diffusion + equivariant graph networks + conditional random field sequence design	NR	NR	NR	primary_generator
Conditioner energies / model likelihood	NR	NR	NR	NR	integrated_scorer
Internal structural generation; not an independent validator	NR	NR	NR	NR	integrated_structure_predictor
Independent folding/complex and 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
Chroma structure diffusion + sequence module	primary_generator	integrated	required	generation
Conditioner energies / model likelihood	integrated_scorer	integrated	optional	post_hoc
Internal structural generation; not an independent validator	integrated_structure_predictor	integrated	optional	post_hoc
Independent folding/complex and 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 protein evidence	Backbone diffusion + equivariant graph networks + conditional random field sequence design	5. prospective functional validation	Early biophysical and crystallographic validation; exact count: NR	NR	General protein evidence	[MATRIX] landscape matrix 2026-08-13	Не читать как полный разбор primary paper.