人工智能驱动合成生物学研究

兰晶岗; 傅雄飞; 汪小我; 张先恩

doi:10.12211/2096-8280.2026-007

您当前的位置：

首页 >

文章列表页 >

人工智能驱动合成生物学研究

观点 | 更新时间：2026-05-07

- 人工智能驱动合成生物学研究
- AI + synthetic biology： a paradigm shift in biomanufacturing
- 合成生物学 2026年页码：1-13
- 作者机构：
  
  1.深圳理工大学合成生物学院，广东深圳 518107
  2.中国科学院深圳先进技术研究院，定量合成生物学全国重点实验室，深圳合成生物学创新研究院，广东深圳 518055
  3.清华大学自动化系，合成与系统生物学研究中心，教育部生物信息学重点实验室，北京信息科学与技术国家研究中心，北京 100084
- 作者简介：
  
  兰晶岗，男，深圳理工大学合成生物学院特聘教授。研究方向为计算化学与AI for Science，机器学习方法研发及其在生物系统与物理体系中的跨尺度应用。
  [ "傅雄飞，男，研究员，中国科学院深圳先进技术研究院合成生物学研究所所长，定量合成生物学全国重点实验室副主任，深圳合成生物创新研究院副院长，深圳合成生物学协会会长。主要从事合成生物学、定量生物学、统计物理与复杂系统等多个交叉学科领域，聚焦随机涨落在细菌单细胞与多细胞有序空间结构形成过程中的作用，通过定量理论与合成重构结合，揭示跨尺度涌现性原理。 E-mail：xiongfei.fu@siat.ac.cn" ]
  [ "汪小我，男，博士，教授。主要研究方向为模式识别与机器学习、生物信息学、合成生物学。 E-mail：xwwang@tsinghua.edu.cn" ]
  张先恩，男，深圳理工大学合成生物学院院长、讲席教授，中国科学院生物物理所研究员。从事合成生物学、生物传感和纳米生物学交叉创新研究，并用于解决细胞生物学、病毒学、肿瘤生物学问题。 E-mail：zhangxianen@suat-sz.edu.cn；
- 基金信息：
- DOI：10.12211/2096-8280.2026-007
  中图分类号： Q81
- 收稿：2026-02-25，
  
  修回：2026-04-21，
  
  网络首发：2026-05-07，
- 稿件说明：
移动端阅览
兰晶岗，傅雄飞，汪小我，张先恩. 人工智能驱动合成生物学研究［J］. 合成生物学， 2026， 7. DOI： 10.12211/2096-8280.2026-007

LAN Jinggang， FU Xiongfei， WANG Xiaowo， ZHANG Xian-En. AI + synthetic biology： a paradigm shift in biomanufacturing［J］. Synthetic Biology Journal， 2026， 7. DOI： 10.12211/2096-8280.2026-007
兰晶岗，傅雄飞，汪小我，张先恩. 人工智能驱动合成生物学研究［J］. 合成生物学， 2026， 7. DOI： 10.12211/2096-8280.2026-007 DOI：

LAN Jinggang， FU Xiongfei， WANG Xiaowo， ZHANG Xian-En. AI + synthetic biology： a paradigm shift in biomanufacturing［J］. Synthetic Biology Journal， 2026， 7. DOI： 10.12211/2096-8280.2026-007 DOI：

摘要

人工智能（AI）正在深刻重塑合成生物学的研究范式，使生命系统的设计从经验驱动迈向模型驱动。传统合成生物学依赖突变筛选与试错式优化，难以应对多尺度、高维度、强耦合的生命过程。随着组学数据的爆发式增长、自动化实验平台的普及以及深度学习技术的快速发展，AI 为揭示序列—结构—功能规律、构建可预测生物模型和实现大规模生命设计提供了全新路径。AI 合成生物学已在四个关键层级形成系统化框架：在生物大分子层面，蛋白语言模型与生成式结构模型使从头设计酶、受体与自组装材料成为可能；在基因组层面，深度学习推动突变机制建模、大片段序列生成与谱系动力学推断，为可编程基因组构建奠定基础；在细胞层面，AI 与机理模型结合加速虚拟细胞构建，使细胞行为实现可量化预测；在平台层面，多智能体与自动化“设计-建造-测试-学习”（DBTL）循环支持路径规划、酶功能预测与实验调度的全流程自动化。总体而言，AI 使合成生物学从局部优化走向系统生成，从经验探索走向预测设计，为生命系统的可控重构和生物制造创新提供了核心动力。

Abstract

Artificial intelligence (AI) is profoundly reshaping the research paradigm of synthetic biology

shifting the design of living systems from empirically driven approaches to model-driven ones. Traditional synthetic biology relies on mutagenesis screening and trial-and-error optimization

making it difficult to address multiscale

high-dimensional

and strongly coupled biological processes. With the explosive growth of omics data

the widespread adoption of automated experimental platforms

and the rapid development of deep learning technologies

AI provides a new pathway to uncover sequence–structure–function relationships

build predictive biological models

and enable large-scale design of living systems. AI-driven synthetic biology has established a systematic framework across four key levels: at the biomacromolecular level

protein language models and generative structural models make de novo design of enzymes

receptors

and self-assembling materials possible; at the genomic level

deep learning advances modeling of mutational mechanisms

large-fragment sequence generation

and inference of phylogenetic dynamics

laying the foundation for programmable genome construction; at the cellular level

the integration of AI with mechanistic models accelerates virtual cell development

enabling quantitatively predictive descriptions of cellular behavior; at the platform level

multi-agent systems and automated "design–build–test–learn" (DBTL) cycles support end-to-end automation of pathway planning

enzyme function prediction

and experimental scheduling. Overall

AI is moving synthetic biology from local optimization to system-level generation

and from empirical exploration to predictive design

providing a core driving force for the controllable reprogramming of living systems and innovation in biomanufacturing.

关键词

Keywords

references

“中国学科及前沿领域发展战略研究”项目组 . 中国合成生物学2035发展战略 [M ] . 北京 : 科学出版社 , 2023 .

Chinese Disciplines and Frontier Fields Development Strategy Research Project Group . Development strategy of synthetic biology 2035 in China [M ] . Beijing : Science Press , 2023 .

张先恩 . 十大措施合力探索合成生物学发展 “深圳模式” [J ] . 中国科学院院刊 , 2024 , 39 ( 9 ): 1574 - 1582 .

ZHANG X E . Ten measures to jointly explore Shenzhen model of synthetic biology development [J ] . Bulletin of Chinese Academy of Sciences , 2024 , 39 ( 9 ): 1574 - 1582 .

JUMPER J , EVANS R , PRITZEL A , et al . Highly accurate protein structure prediction with AlphaFold [J ] . Nature , 2021 , 596 ( 7873 ): 583 - 589 .

BAKER D , SALI A . Protein structure prediction and structural genomics [J ] . Science , 2001 , 294 ( 5540 ): 93 - 96 .

合成生物学发展战略研究组 . 合成生物学路线图-2030: 驱动下一代生物制造的引擎 [M ] . 北京 : 科学出版社 , 2024 .

Strategic Research Group for the Development of Synthetic Biology . Roadmap of synthetic biology-2030: engine driving the next generation bio-manufacturing [M ] . Beijing : Science Press , 2024 .

李玉娟 , 傅雄飞 , 张先恩 . 合成生物学发展脉络概述 [J ] . 中国生物工程杂志 , 2024 , 44 ( 1 ): 52 - 60 .

LI Y J , FU X F , ZHANG X E . A brief overview of synthetic biology [J ] . China Biotechnology , 2024 , 44 ( 1 ): 52 - 60 .

ZHOU C , LI H C , WANG J X , et al . De novo designed voltage-gated anion channels suppress neuron firing [J ] . Cell , 2025 , 188 ( 26 ): 7495 - 7511.e21 .

XU C F , LU P L , GAMAL EL-DIN T M , et al . Computational design of transmembrane pores [J ] . Nature , 2020 , 585 ( 7823 ): 129 - 134 .

ZHU J Y , LIANG M F , SUN K , et al . De novo design of transmembrane fluorescence-activating proteins [J ] . Nature , 2025 , 640 ( 8057 ): 249 - 257 .

SHEFFLER W , YANG E C , DOWLING Q , et al . Fast and versatile sequence-independent protein docking for nanomaterials design using RPXDock [J ] . PLoS Computational Biology , 2023 , 19 ( 5 ): e1010680 .

LI Z , WANG S Z , NATTERMANN U , et al . Accurate computational design of three-dimensional protein crystals [J ] . Nature Materials , 2023 , 22 ( 12 ): 1556 - 1563 .

LIU Y F , WANG S , DONG J X , et al . De novo protein design with a denoising diffusion network independent of pretrained structure prediction models [J ] . Nature Methods , 2024 , 21 ( 11 ): 2107 - 2116 .

LIU Y F , ZHANG L , WANG W L , et al . Rotamer-free protein sequence design based on deep learning and self-consistency [J ] . Nature Computational Science , 2022 , 2 ( 7 ): 451 - 462 .

DING Y , LUO X , GUO J S , et al . Identif ication of gut microbial bile acid metabolic enzymes via an AI-assisted pipeline [J ] . Cell , 2025 , 188 ( 21 ): 6012 - 6027.e20 .

JIN Q H , WANG Y K , CHEN D C , et al . De novo design of small molecule-regulated protein oligomers [J ] . Science , 2026 , 391 ( 6780 ): eady6017 .

SUN J Y , ZHU T , CUI Y L , et al . Structure-based self-supervised learning enables ultrafast protein stability prediction upon mutation [J ] . Innovation , 2025 , 6 ( 1 ): 100750 .

TANG Y , TIAN X Y , WANG M , et al . The β-d- manno -heptoses are immune agonists across Kingdoms [J ] . Science , 2024 , 385 ( 6709 ): 678 - 684 .

CHEN Y C , SUN J Y , SHI K L , et al . Glycolysis-compatible urethanases for polyurethane recycling [J ] . Science , 2025 , 390 ( 6772 ): 503 - 509 .

ZHANG C X , LIU Q C , FREDDOLINO L . StarFunc: fusing template-based and deep learning approaches for accurate protein function prediction [J ] . Genomics , Proteomics & Bioinformatics, 2026 : qzag018 .

MA R Z , ZHANG Z Y , WANG Z C , et al . RiboFlow: conditional de novo RNA co-design via synergistic flow matching [C/OL ] The Thirty-ninth Annual Conference on Neural Information Processing Systems , 2025 [ 2026-02-26 ] . https://neurips.cc/virtual/2025/loc/san-diego/poster/117145 https://neurips.cc/virtual/2025/loc/san-diego/poster/117145 .

FANG Y Y , DENG S Y , LI C . A generalizable deep learning framework for inferring fine-scale germline mutation rate maps [J ] . Nature Machine Intelligence , 2022 , 4 ( 12 ): 1209 - 1223 .

XU K , YU Z Y , SUN C Z , et al . ChromInSight: revealing DNA double-strand breaks through chromatin structural insights with an interpretable graph neural network framework [J ] . Advanced Science , 2025 , 12 ( 36 ): e04571 .

LI X W , LIANG Z , GUO , Z T , et al . Leveraging large language models for metabolic engineering design [PP/OL ] . bioRxiv ( 2024-09-13 )[ 2025-02-26 ] . https://doi.org/10.1101/2024.09.09.612023 https://doi.org/10.1101/2024.09.09.612023 .

HU J L , AMOR D R , BARBIER M , et al . Emergent phases of ecological diversity and dynamics mapped in microcosms [J ] . Science , 2022 , 378 ( 6615 ): 85 - 89 .

YANG W , HICKS D R , GHOSH A , et al . Design of high-affinity binders to immune modulating receptors for cancer immunotherapy [J ] . Nature Communications , 2025 , 16 : 2001 .

ZENG Z S , XU R F , GUO J , et al . Accelerating functional protein discovery with GPT models: Antimicrobials and enzymes [J ] . The Innovation Life , 2025 , 3 ( 2 ): 100133 .

ZHENG B , LU Z J , WANG S C , et al . Computational design of superstable proteins through maximized hydrogen bonding [J ] . Nature Chemistry , 2026 , 18 ( 2 ): 364 - 373 .

LU W , ZHANG J X , HUANG W F , et al . DynamicBind: predicting ligand-specific protein-ligand complex structure with a deep equivariant generative model [J ] . Nature Communications , 2024 , 15 : 1071 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

生物制造标准体系建设的现状、问题与建议

解脂耶氏酵母底盘细胞的工程改造及应用

基于人工智能和计算生物学的合成生物学元件设计

“可折叠性”在酶智能设计改造中的应用研究——以AlphaFold2为例

微生物组生物合成基因簇发掘方法及应用前景