Главная / Инструменты / Makowski et al. affinity–specificity co-optimization

Makowski et al. affinity–specificity co-optimization

Makowski emibetuzumab co-optimization

Peer-reviewed Prospective flag Direct antibody flag

Summary card

Поле	Значение
Год / version	2022 / NR
Last verified	2026-08-13
Что генерирует	Combinatorial CDR mutation sets predicted to trade affinity against nonspecific binding
Форматы	Therapeutic antibody (emibetuzumab) variants
De novo или optimization	Optimizes existing antibody
DOI	10.1038/s41467-022-31457-3
Paper	https://doi.org/10.1038/s41467-022-31457-3
GitHub / repo	NR
Service / docs	NR
Лицензия	NR
Доступность / commercial-use	Workflow/paper; not a packaged tool

Задача и generated object

Used in these tasks: Combinatorial mutation design, Multi-objective optimization, Оптимизация specificity, Affinity maturation, Оптимизация polyspecificity

Part of these workflows: Few-shot / labelled mutation optimization

Поле	Значение
Required input	Parent antibody, predefined CDR positions, sorted library/deep-sequencing labels
Conditioning	Eight positions in three heavy-chain CDRs in reported campaign
Fixed	NR
Editable	NR
Generated	Combinatorial CDR mutation sets predicted to trade affinity against nonspecific binding
Single mutations	Yes
Combinatorial mutations	Yes
Multi-objective	Yes—affinity and specificity/nonspecific binding

Architecture and operating principle

Author-reported (matrix)

Поле	Значение
Base architecture	NR
Model type	6 Mutation generator / optimization workflow
Representation	NR
Generation unit	Combinatorial CDR mutation sets predicted to trade affinity against nonspecific binding
Conditioning	Eight positions in three heavy-chain CDRs in reported campaign
Output	Combinatorial CDR mutation sets predicted to trade affinity against nonspecific binding
Sampling	NR
Internal scorer	Affinity- and specificity-correlated supervised models
Structure generation	No structure 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

Eight positions in three heavy-chain CDRs in reported campaign

Generation mechanism

Supervised models on display/deep-sequencing labels + Pareto-front proposal

Sampling: NR

Training objectives

Losses: NR

Mathematical details: NR

Integrated evaluation

Affinity- and specificity-correlated supervised models

Architectural limitations

Requires campaign-specific labelled data and library construction.
not a transferable pretrained generator.
one antibody case study.

Uses these architecture components: Multi-objective optimization, Few-shot mutation optimization

Generator versus evaluation

Component	Role	Architecture class	Integrated or external	Required or optional	Influences generation or post hoc	Runtime included or excluded
Supervised search/proposal over mutational space	primary_generator	Supervised models on display/deep-sequencing labels + Pareto-front proposal	integrated	required	generation	included
Affinity- and specificity-correlated supervised models	integrated_scorer	NR	integrated	optional	post_hoc	unknown
Experimental display/SPR and developability panel	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 prospective binding/specificity validation
Retrospective	NR
Prospective binding	Co-optimized variants beyond sampled library; affinity gains trade against specificity
Prospective function	Co-optimized variants beyond sampled library; affinity gains trade against specificity
Developability	Co-optimized variants beyond sampled library; affinity gains trade against specificity
Tested design count	~10^7 variant library; 4,000 sequences used for modeling; 125 clones characterized
Success criterion	Co-optimized variants beyond sampled library; affinity gains trade against specificity
Independent validation	NR
Summary	Co-optimized variants beyond sampled library; affinity gains trade against specificity

Limitations and failure modes

Requires campaign-specific labelled data and library construction.
not a transferable pretrained generator.
one antibody case study.

Deployment: Limited custom internal pilot

Reproducibility, license, commercial use

Поле	Значение
Repo	NR
License	NR
Commercial-use status	Workflow/paper; not a packaged tool

Alternative tools: NR

Complementary: NR

Related: NR

Not directly comparable with

Efficient evolution with ESM-1v, FrameDiff, RFdiffusion, RFdiffusion All-Atom (RFdiffusionAA), 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

Co-optimization of therapeutic antibody affinity and specificity using machine learning models of sequence-activity data

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

Review queue

Official repository/service not verified.

Детальный source-grounded разбор

1. Что именно делает инструмент

[MATRIX]Вход: Parent antibody, predefined CDR positions, sorted library/deep-sequencing labels

[MATRIX]Generated object: Combinatorial CDR mutation sets predicted to trade affinity against nonspecific binding

[MATRIX]Фиксируется: NR. Изменяется: NR.

[MATRIX]De novo или parent-based: Optimizes existing antibody.

2. Архитектура в одном абзаце

[MATRIX]Supervised models on display/deep-sequencing labels + Pareto-front proposal

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: Combinatorial CDR mutation sets predicted to trade affinity against nonspecific binding. Structure generation: No structure 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): Co-optimized variants beyond sampled library; affinity gains trade against specificity

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 prospective binding/specificity validation. Prospective binding: Co-optimized variants beyond sampled library; affinity gains trade against specificity. Function: Co-optimized variants beyond sampled library; affinity gains trade against specificity. Developability: Co-optimized variants beyond sampled library; affinity gains trade against specificity.

Requires campaign-specific labelled data and library construction.
not a transferable pretrained generator.
one antibody case study.

15. Reproducibility

NRРепозиторий: NR. Лицензия: NR. Commit: NR, пока страница не имеет отдельного analysis YAML.

Модульная схема

Модуль	Класс модели	Вход	Выход	Обучается или frozen	Generator или evaluator
Supervised search/proposal over mutational space	Supervised models on display/deep-sequencing labels + Pareto-front proposal	NR	NR	NR	primary_generator
Affinity- and specificity-correlated supervised models	NR	NR	NR	NR	integrated_scorer
Experimental display/SPR and developability panel	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
Supervised search/proposal over mutational space	primary_generator	integrated	required	generation
Affinity- and specificity-correlated supervised models	integrated_scorer	integrated	optional	post_hoc
Experimental display/SPR and developability panel	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
Co-optimized variants beyond sampled library; affinity gains trade against specificity	Supervised models on display/deep-sequencing labels + Pareto-front proposal	5. prospective functional validation	~10^7 variant library; 4,000 sequences used for modeling; 125 clones characterized	NR	Co-optimized variants beyond sampled library; affinity gains trade against specificity	[MATRIX] landscape matrix 2026-08-13	Не читать как полный разбор primary paper.