Главная / Инструменты / ProGen / ProGen2

ProGen / ProGen2

ProGen ProGen2

Peer-reviewed Prospective flag

Summary card

Поле	Значение
Год / version	2023 / NR
Last verified	2026-08-13
Что генерирует	Amino-acid sequence
Форматы	Monomer sequences; family-specific proteins
De novo или optimization	Mostly de novo sequence; fine-tuning can localize a family
DOI	10.1038/s41587-022-01618-2
Paper	https://doi.org/10.1038/s41587-022-01618-2
GitHub / repo	https://github.com/salesforce/progen
Service / docs	NR
Лицензия	BSD-3-Clause
Доступность / commercial-use	Code + checkpoints

Задача и generated object

Used in these tasks: Генерация белковых sequences

Part of these workflows: Sequence-first de novo generation

Поле	Значение
Required input	Optional family/function tags; optional seed/prefix
Conditioning	Tags, prefix; no native 3D constraints
Fixed	NR
Editable	NR
Generated	Amino-acid sequence
Single mutations	Indirect
Combinatorial mutations	Yes (full sequence)
Multi-objective	No explicit Pareto loop

Architecture and operating principle

Author-reported (matrix)

Поле	Значение
Base architecture	NR
Model type	1 Sequence generator
Representation	NR
Generation unit	Amino-acid sequence
Conditioning	Tags, prefix; no native 3D constraints
Output	Amino-acid sequence
Sampling	NR
Internal scorer	Sequence likelihood; paper used external computational triage
Structure generation	No dedicated structure/complex predictor

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

Tags, prefix; no native 3D constraints

Generation mechanism

Autoregressive Transformer; control tags / family conditioning

Sampling: NR

Training objectives

Losses: NR

Mathematical details: NR

Integrated evaluation

Sequence likelihood; paper used external computational triage

Architectural limitations

Sequence plausibility is not structural or functional proof.
conditioning vocabulary is coarse.
family transfer requires care.

Uses these architecture components: Protein language models, Autoregressive generation

Generator versus evaluation

Component	Role	Architecture class	Integrated or external	Required or optional	Influences generation or post hoc	Runtime included or excluded
ProGen autoregressive LM	primary_generator	Autoregressive Transformer; control tags / family conditioning	integrated	required	generation	included
Sequence likelihood; paper used external computational triage	integrated_scorer	NR	integrated	optional	post_hoc	unknown
Structure prediction + task-specific functional assays	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	4 prospective functional validation
Retrospective	Functional artificial sequences across several lysozyme families; exact aggregate count: NR here
Prospective binding	Prospective enzymatic function reported; not antibody evidence
Prospective function	Prospective enzymatic function reported; not antibody evidence
Developability	NR
Tested design count	Functional artificial sequences across several lysozyme families; exact aggregate count: NR here
Success criterion	Prospective enzymatic function reported; not antibody evidence
Independent validation	NR
Summary	Prospective enzymatic function reported; not antibody evidence

Limitations and failure modes

Sequence plausibility is not structural or functional proof.
conditioning vocabulary is coarse.
family transfer requires care.

Deployment: Limited internal pilot

Reproducibility, license, commercial use

Поле	Значение
Repo	https://github.com/salesforce/progen
License	BSD-3-Clause
Commercial-use status	Code + checkpoints

Alternative tools: NR

Complementary: NR

Related: NR

Not directly comparable with

AntiFold, ESM-IF1, FrameDiff, LigandMPNN, ProteinMPNN, RFdiffusion, RFdiffusion All-Atom (RFdiffusionAA)

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

Large language models generate functional protein sequences across diverse families

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

Детальный source-grounded разбор

1. Что именно делает инструмент

[MATRIX]Вход: Optional family/function tags; optional seed/prefix

[MATRIX]Generated object: Amino-acid sequence

[MATRIX]Фиксируется: NR. Изменяется: NR.

[MATRIX]De novo или parent-based: Mostly de novo sequence; fine-tuning can localize a family.

2. Архитектура в одном абзаце

[MATRIX]Autoregressive Transformer; control tags / family conditioning

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: Amino-acid sequence. Structure generation: No dedicated structure/complex predictor.

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): Prospective enzymatic function reported; not antibody 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): 4 prospective functional validation. Prospective binding: Prospective enzymatic function reported; not antibody evidence. Function: Prospective enzymatic function reported; not antibody evidence. Developability: NR.

Sequence plausibility is not structural or functional proof.
conditioning vocabulary is coarse.
family transfer requires care.

15. Reproducibility

NRРепозиторий: https://github.com/salesforce/progen. Лицензия: BSD-3-Clause. Commit: NR, пока страница не имеет отдельного analysis YAML.

Модульная схема

Модуль	Класс модели	Вход	Выход	Обучается или frozen	Generator или evaluator
ProGen autoregressive LM	Autoregressive Transformer; control tags / family conditioning	NR	NR	NR	primary_generator
Sequence likelihood; paper used external computational triage	NR	NR	NR	NR	integrated_scorer
Structure prediction + task-specific functional assays	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
ProGen autoregressive LM	primary_generator	integrated	required	generation
Sequence likelihood; paper used external computational triage	integrated_scorer	integrated	optional	post_hoc
Structure prediction + task-specific functional assays	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
Prospective enzymatic function reported; not antibody evidence	Autoregressive Transformer; control tags / family conditioning	5. prospective functional validation	Functional artificial sequences across several lysozyme families; exact aggregate count: NR here	NR	Prospective enzymatic function reported; not antibody evidence	[MATRIX] landscape matrix 2026-08-13	Не читать как полный разбор primary paper.