摘要

1988 年前后,de novo 标记的是不从天然序列改起、而按物理原则写出可折叠链。Korendovych 与 DeGrado 将此后三十多年分为手工模型、物理化学计算、片段与信息学组装;Woolfson 使用最小、理性、计算三条时间轴;Huang、Boyken 与 Baker 在 2016 年把原子级结构设计写成可重复实践。2024 年诺贝尔化学奖把计算设计(Baker)与结构预测(Hassabis、Jumper)分开授奖。预测精度不等于设计成功率。2022–2025 年的实验论文把二者接到同一管线:RFdiffusion 或 BindCraft 给出骨架,ProteinMPNN 或 LigandMPNN 填写序列,AlphaFold2、RoseTTAFold 或 ESMFold 做自洽过滤,再进入表达与晶体或电镜。

人物清点以 RosettaCommons 实验室名录(2026-09-14)为主,并纳入名录未覆盖的预测、生成与产业第一作者,共 192 人。分期看实验产物:NMR 或晶体上的螺旋束,Top7 级新拓扑,再往后是可筛选的功能蛋白。Baker/IPD 出软件和博士后。DeGrado 出化学原则。Correia 与 Fleishman 出功能化。上海交大、中科大、中科院计算所、清华、北大构成数据驱动与功能工程轴(SCUBA/ABACUS、CarbonDesign、TopoDiff、Venus、GeoEvoBuilder)。西湖大学的曹龙兴、卢培龙、陈子博、原发杰把湿实验留在国内,方法引用仍指向 IPD 仓库。结构预测作者多数不做湿实验设计。BioEmu、ESMDynamic 和开关蛋白把目标从单点能量极小扩到构象分布和相对自由能。

192
清点人物(去重)
79
Rosetta / 物理计算相关
47
预测 + 生成模型作者
20
中国独立节点(名录内)

1. 引言与研究问题

1997 年 Dahiyat 与 Mayo 用死端消除写出 FSD-1,合成了一条 28 残基、不再依赖锌离子的折叠单元。2003 年 Kuhlman、Dantas、Ireton、Varani、Stoddard 与 Baker 给出 Top7 晶体,新拓扑与设计模型在原子尺度吻合。2023 年 Watson、Juergens、Bennett 等人把 RoseTTAFold 改成扩散生成器 RFdiffusion,并在 Nature 报告实验表征的新骨架与 binder。中间二十多年,实验室叙事经常是大量计算、少量晶体。

已有综述把结构写清楚了,公开可检索的人事迁徙表仍然少。工具栈跟着人事走:ProteinMPNN 仓库在 Dauparas 名下,BindCraft 的实验闭环在 Correia 组,西湖的小分子响应寡聚体(曹龙兴组,Science 2026)使用同一套 diffusion / MPNN / AF 输入文件语言。

RQ1

方法史如何分期,每一期留下的可检验产物是什么——晶体吻合、功能周转、还是只是 in silico 自洽?

RQ2

至少 100 位核心人物如何通过师生、合作、公共代码与创业形成可追踪交互?哪些边是公开名录,哪些仍未证实?

RQ3

2020 年后 AlphaFold2 / RoseTTAFold / ESM 如何改写设计闭环?哪些主张有湿实验结构或功能,哪些仍停在预印本?

2. 方法与证据边界

覆盖五条公开文献线:Rosetta / IPD 计算设计,DeGrado / Woolfson 化学理性设计,CASP / DeepMind / ESM 结构预测,RFdiffusion、BindCraft、Chroma 生成管线,以及中国与欧洲独立实验室。主名录来自 RosettaCommons Labs 2026-09-14 页面。补入人物限于里程碑论文、研究所软件页或公司官方页可核对者。

纳入公开发表过 de novo 或计算蛋白质设计的人,以及提供设计闭环必需工具(预测、逆折叠、MSA、评估)的工具负责人。仅做天然蛋白改造、无计算设计主张者(如阿诺德的定向进化)标为相邻。师生边若无实验室主页或论文致谢,表中写名录在列或合作可见,不补造导师。

证据强度:同行评审实验结构(晶体、NMR、电镜)最高;功能测定次之;in silico 序列恢复与 AF2 pLDDT 只说明过滤器。预印本(AlphaProteo、部分 RF3、部分公司技术白皮书)单独标注。

3. 分类轴

横轴是产物:折叠原则,新拓扑,结合与催化,可编程开关与组装。纵轴是机构传统:UCSF 化学、Caltech ORBIT、西雅图 IPD、洛桑 EPFL、魏茨曼、西湖、北京–清华–中科院–交大、DeepMind/Meta。de novo 抗体与酶的实验成功率低于 minibinder。膜蛋白设计人数少于可溶螺旋束。Helixon、字节 Protenix、分子之心的湿实验第一作者公开程度不规则,表中只收能核对的人名。

4. 中国两条学术体系

西雅图 IPD 与伦敦 DeepMind 解释工具从哪里来。中国国内还有两条可以按机构和产物分开写的线。一条以北京为重心,沿中科院与高校计算结构生物学铺开,接到上海交大的功能工程。一条以西湖大学为实验闭环,人员来自 Baker 实验室,课题转向膜蛋白、开关与配体门控组装。

上海交大–中科大–中科院–清华–北大

中国科学技术大学施蕴渝长期做计算结构生物学。刘海燕与陈泉在此基础上建立数据驱动设计:ABACUS 给定主链写序列,SCUBA 在序列未定时搜索可设计主链。2022 年 Nature 报告 9 个晶体结构,其中多种拓扑在天然蛋白中未见。2024 年 Nature Methods 的 SCUBA-D 把主链生成改成不依赖预训练结构预测网络的扩散模型,并给出单体与 Ras 结合蛋白的晶体。

中科院计算技术研究所卜东波、张海仓的 CarbonDesign(Nature Machine Intelligence 2024)把 AlphaFold 式的三角更新改成 Inverseformer 做逆折叠,并在 CASP15、CAMEO 与 RFdiffusion 骨架上报告序列恢复。CarbonNovo(ICML 2024)尝试结构与序列联合生成。清华大学龚海鹏与北京生命科学研究所徐春福的 TopoDiff(Nature Machine Intelligence 2025)用 32 维全局几何潜变量控制扩散骨架,并报告新颖 β 折叠的实验验证。

北京大学来鲁华组的 GeoEvoBuilder(2025)在一轮设计中同时抬高活性与热稳定性,优化的是功能序列,不是全新折叠。上海交通大学洪亮组的 Venus 系列把蛋白质语言模型接到酶挖掘与多目标定向进化,并配套表达纯化检测一体机;公开材料强调产业酶。马剑竹(清华 AIR,博士导师许锦波)做抗体、肽与口袋生成。方法论文可以不经过 RosettaCommons 人事。湿实验规模因组而异:SCUBA 与 TopoDiff 有晶体,Venus 与 GeoEvoBuilder 更靠近酶与活性。

西湖大学

西湖生命科学学院把 Baker 博士后变成独立 PI。卢培龙做跨膜孔道、电压门控阴离子通道与跨膜荧光激活蛋白(Nature 2020 及后续 Cell/Nature 工作)。曹龙兴做 minibinder 与配体门控寡聚体(Science 2026),以及非天然氨基酸光开关。陈子博做蛋白电路与可编程自组装。原发杰做 SaProt、ProTrek 等结构感知语言模型。方法仍引用 RFdiffusion、ProteinMPNN、AlphaFold。实验、膜体系和开关功能在西湖完成。与北方轴的差别是人事来源和产物类型:这边是 Baker 博士后,膜、开关、组装更多。

5. 方法史:实验产物换了四次

1961–1988 · 前提

Anfinsen 把折叠写成序列的热力学函数。Levinthal 指出随机搜索不可能。Ponder 与 Richards 的旋转异构体近似,后来成为 Mayo 死端消除与 Rosetta 侧链打包的共同祖先。此期没有 de novo 蛋白,只有问题的提法。

1987–2000 · 理性与最小设计

DeGrado、Regan 的四螺旋束把“de novo”变成可合成对象。Hecht 的二元密码证明疏水/亲水图案足以得到类蛋白。Harbury、Alber、Kim 在 GCN4 上用 packing 解释寡聚状态。Woolfson 把 coiled-coil 规则做成可组合零件。可检验产物:NMR/晶体上的螺旋束,功能仍然稀少。

1997–2016 · 全原子计算

Dahiyat/Mayo 1997 Science:FSD-1。Kuhlman/Baker 2003 Science:Top7,新拓扑、原子级吻合。2008 年 Baker 组报告逆羟醛酶等计算酶,活性相对天然酶仍弱。2013 年前后 Koga 夫妇与 Huang 把蓝图规则写成可批量的单体折叠。2016 年 Huang、Boyken、Baker 在 Nature 写 The coming of age of de novo protein design。产物是多样拓扑的高分辨率结构。功能设计仍贵。

2020–2026 · 预测器进入设计环

AlphaFold2(Jumper 等,Nature 2021)与 RoseTTAFold(Baek 等,Science 2021)把结构预测变成公共过滤器。ProteinMPNN(Dauparas 等,Science 2022)序列恢复 52.4% 对 Rosetta 32.9%,并救活失败设计。RFdiffusion(Watson 等,Nature 2023)从噪声雕骨架。Chroma(Generate)与 ESMFold(Lin 等)提供并行路径。BindCraft(Pacesa 等,Nature 2025)把 AF2 梯度焊进 binder 管线。LigandMPNN(Dauparas 等,Nature Methods 2025)把配体原子写进逆折叠。2024 年诺贝尔化学奖一半给计算设计,一半给结构预测。

Kortemme 2024 年 Cell 述评写的是:新折叠与高阶组装已经能在实验室里做出来,可调构象、极性分子识别、从零写细胞信号还难。Chu、Lu 与 Huang 2024 年 Nat Biotechnol 把同一缺口写成 Sparks of function:结构会了,功能刚开始。

6. 热力学目标与动力学增强

经典 de novo 设计优化热力学:序列在目标骨架上的能量是否低于竞争构象。Rosetta 全原子力场、Mayo 的死端消除、SCUBA 的神经网络能量项都在做这件事。BioEmu(Lewis、Noé 等,Science 2025)把同一目标改写成可采样的平衡系综:单块 GPU 每小时给出数千个统计独立结构,相对自由能误差相对毫秒级 MD 与实验稳定性约 1 kcal/mol,并覆盖隐蔽口袋、局部去折叠与结构域重排。训练目标是 Boltzmann 分布。

动力学正在进入设计对象。ESMDynamic(Kleiman 等,Nature Communications 2026)从单序列预测残基接触的动态占据,用 mdCATH 校准,推理是一次前向传播。AlphaFlow 与 ESMFlow 用生成模型抽构象。开关蛋白、配体门控寡聚体、光响应非天然氨基酸(曹龙兴组)至少需要两个稳态及其转换。Kortemme 把可编程构象写成下一阶段工程原则。龚海鹏组同时维护分子动力学算法,与 TopoDiff 的静态骨架生成并行。

开/关速率、滞留时间、诱导契合路径还没有写进损失函数。现有管线仍用 AF2 pLDDT 与晶体 RMSD 验收单点结构。热力学系综给出 ΔG,接触动力学给出涨落,两者都不能替代结合与催化测定。要把切换时间写进设计目标,BioEmu 类自由能和 ESMDynamic 类接触图需要接到 RFdiffusion 或 SCUBA-D 的生成环里,不能只做事后分析。

对接与旋转异构体场:RIFDock 一类方法

在扩散生成成为蛋白–蛋白 minibinder 默认选项之前,主流是把已有骨架对接到靶标界面。RIFDock(Rotamer Interaction Field docking;Coventry 实现,Cao、Coventry 等 Nature 2022)先在靶标表面预计算孤立侧链的有利位置(RIFGen),再把小蛋白骨架在六维刚体空间里对进这个场,最后用 Rosetta 填序列。靶标骨架保持固定。同一论文报告了对 12 类形状各异靶标的 minibinder,是曹龙兴组后续工作的直接前身。

2025–2026 年这类方法没有消失,位置变了。BindCraft(Pacesa 等,Nature 2025)仍把 RIFdock 与 RFdiffusion 并列为“靶标骨架固定”的对照:BindCraft 重预测 binder–靶标复合物,RIFDock 不重预测。Glasscock 等设计序列特异 DNA 结合蛋白的公开管线(配套脚本 2025)仍把 RIFgen/RIFdock 放在第一步,再接 LigandMPNN、部分 RFdiffusion inpainting 与 AF2 过滤。小分子结合口袋上,RIF 对接给出的姿态仍可被 FastDesign 或 LigandMPNN 救活。仓库 rifdock/rifdock 仍要求 Rosetta 3.9 或 bcov/stable1,没有并入 Rosetta 主干。

仍在用的对接族还有 RosettaDock、ZDOCK(抗体与免疫识别,Pierce 等),以及 MaSIF 表面指纹后接种子对接(Correia,后接到 BindCraft)。蛋白–蛋白全新 binder 现在默认走 RFdiffusion 热点条件生成或 BindCraft。RIFDock 仍出现在 DNA 结合、小分子、已有支架库穷举,以及与扩散结果交叉验证的场合。

7. 人物网络与生涯迁徙

枢纽不止诺奖三人。Baker 出软件与博士后。DeGrado 出化学原则与金属、膜蛋白。Mayo 出自动化组合优化,经 Dahiyat 进入抗体公司。Correia 把 MaSIF 与 BindCraft 做成洛桑功能设计。DeepMind、Meta/ESM、Steinegger、Noé/BioEmu 几乎不养湿实验设计组,但决定过滤阈值。中国侧另有刘海燕/陈泉、来鲁华、龚海鹏/徐春福、卜东波/张海仓、洪亮,以及西湖曹/卢/陈。

下图为枢纽一阶邻域(pyvis 内嵌)。只含可核对生涯边,不含 Commons 名录引力。

8. 可交互网络组织图(pyvis)

网络用 pyvis 生成,vis.js 打进页面。人物网只画可核对生涯边。方法网覆盖原则、力场、生成、过滤、RIFDock、系综。人物×方法把工具接到人。全图另含 Commons 名录引力(虚线,不升格为博士后)。可拖拽、缩放、邻域高亮,右上菜单按组过滤。

可核对生涯边 85 条;名录引力 107 条;方法衍生 48 条;人—方法 48 条。源数据 denovo_network.json

Baker/IPD DeGrado 化学 Mayo / 物理计算 结构预测 ESM / 生成 中国节点 产业 方法节点

可核对的迁徙模式

9. 192 人清点表

下表以姓氏排序。RosettaCommons 名录中的相邻领域(基因组浏览器、神经工具、RNA 设计、公民科学游戏)予以保留并打上“相邻领域”,避免把 Foldit 或 UCSC Genome Browser 写成 de novo 设计实验室。检索框与标签只过滤表格,不改正文。

#姓名机构方法 / 角色生涯交互类别
1Josh AbramsonGoogle DeepMindAlphaFold 3 共折叠2024 Nature;配体/核酸进入同一模型结构预测来源
2Gustaf AhdritzHarvard / OpenFoldOpenFold 实现AlQuraishi 合作结构预测来源
3Woody AhernUniversity of Washington IPDRFdiffusion / RFAA / RFdiffusion2Baker 实验室全原子与酶脚手架;公开材料中的 RFdiffusion 工程负责人之一生成设计来源
4Mohammed AlQuraishiColumbia UniversityOpenFold 可复现 AF2把闭源预测变成可训练的开源栈结构预测来源
5Namrata AnandStanford Huang lab学习到的序列势、TIM barrel 序列Possu Huang 实验室;晶体验证的逆折叠生成设计来源
6Ingemar AndreLund UniversityProtein modeling, symmetric protein complexes, self-assembly, computational design, structural biology.Baker 博士后;Lund 独立RosettaCommons来源
7Christian Anfinsen 克里斯蒂安·安芬森NIH(历史)序列决定结构的热力学前提1972 诺贝尔化学奖;设计问题的物理学前提,本人不做 de novo 设计前提/评估来源
8Ivan AnishchenkoUniversity of Washington IPDtrRosetta / 幻觉 / ProteinMPNN 合作Baker 实验室深度学习主线生成设计来源
9Frances Arnold 弗朗西斯·阿诺德Caltech定向进化(2018 诺贝尔)相邻:不从头写骨架,但常与 de novo 设计串成“设计+进化”相邻领域来源
10Mihai AzoiteiDuke UniversityVaccines, protein engineering, biosensors, influenza, HIV, coronaviruses, B-cell signaling.RosettaCommons 实验室名录RosettaCommons来源
11Minkyung Baek 白玟暻Seoul National UniversityRoseTTAFoldBaker 合作;开源三轨网络,后独立于首尔大学结构预测来源
12Chris BahlAI ProteinsDe novo protein design, peptide design, protein interactions, protein engineering, biologic drugs, drug discovery, high-throughput science, laboratory automation.RosettaCommons 实验室名录产业/创业来源
13David Baker 戴维·贝克University of WashingtonDe novo protein design, hyperstable proteins, vaccine design, RFdiffusion, protein structure, enzyme catalysis, self-assembling nanomaterials, drug delivery, Top7, Nobel Prize.RosettaCommons 实验室名录RosettaCommons来源
14Patrick BarthEPFL LausanneTransmembrane proteins, ligand-receptor interactions, signaling pathways, computational modeling, protein design, membrane biology.RosettaCommons 实验室名录RosettaCommons来源
15Nathaniel BennettUniversity of Washington IPDRFdiffusion / RFantibodyBaker 实验室 binder 与抗体设计生成设计来源
16Gaurav BhardwajUniversity of WashingtonPeptide drug design, computational modeling, antibiotics, antivirals, anti-cancer peptides, cellular permeability, oral bioavailability, machine learning.RosettaCommons 实验室名录RosettaCommons来源
17Surge BiswasNabla Bio生成式抗体/蛋白设计平台与 Sam Sinai;产业侧序列生成产业/创业来源
18Sinisa BjelicLinnaeus UniversityMolecular modeling, protein design, small-molecule therapeutics, ROSETTA, molecular dynamics, docking, biochemistry, drug discovery.RosettaCommons 实验室名录RosettaCommons来源
19Richard BonneauPrescient Design, Genentechcomputational protein designRosettaCommons 实验室名录产业/创业来源
20Ed BoydenMassachusetts Institute of TechnologyNeuroscience, brain mapping, expansion microscopy, optogenetics, synthetic biology, nanofabrication, neural dynamics, therapeutic strategies.RosettaCommons 实验室名录相邻领域来源
21Phil BradleyFred Hutchinson Cancer CenterMolecular recognition, structural modeling, protein-DNA interactions, protein-peptide interactions, Rosetta, molecular modeling algorithms.RosettaCommons 实验室名录RosettaCommons来源
22Ben BrownVanderbilt UniversityProtein dynamics, signal transduction, enzyme catalysis, molecular recognition, AI-driven drug design, induced-fit modeling, virtual screening, disease mutations.RosettaCommons 实验室名录RosettaCommons来源
23Dongbo Bu 卜东波中科院计算技术研究所ProFOLD / CarbonDesign / CarbonNovo计算所生物信息学;序列–结构联合设计中国节点来源
24Christopher BystroffRensselaer Polytechnic InstituteProtein folding, protein design, green fluorescent protein (GFP), fluorescent biosensors, molecular dynamics, biophysical analysis, X-ray crystallography.RosettaCommons 实验室名录RosettaCommons来源
25Longxing Cao 曹龙兴Westlake UniversityProtein design, protein interfaces, protein switches, deep learning, synthetic biology, immunotherapy.Baker 博士后 2017–21;西湖独立 PI中国节点来源
26Quan Chen 陈泉中国科学技术大学SCUBA/ABACUS 实验闭环与刘海燕共同通讯中国节点来源
27Zibo Chen 陈子博Westlake UniversityDe novo protein design, molecular computing, protein circuits, self-assembly, Rosetta, programmable biomolecules, supramolecular structures, cellular functions.RosettaCommons 实验室名录中国节点来源
28Qian Cong 丛倩University of Texas SouthwesternProtein-protein interactions, evolution, genomics, artificial intelligence, bioinformatics, AlphaFold, RoseTTAFold.RosettaCommons 实验室名录中国节点来源
29Seth CooperNortheastern UniversityFoldit, citizen science, crowdsourcing, protein structure prediction, protein design, human-computer collaboration.RosettaCommons 实验室名录相邻领域来源
30Jacob CornETH ZurichGenome editing, gene regulation, DNA repair, hematopoietic stem cells, ubiquitin signaling, cellular biochemistry, translational research.RosettaCommons 实验室名录相邻领域来源
31Bruno CorreiaEPFL LausanneProtein design, functional proteins, therapeutics, vaccines, biosensors, synthetic biology.Baker/Schief 博士后;洛桑独立功能设计枢纽生成设计来源
32Gabriele CorsoMIT / BoltzBoltz / DiffDock 几何生成开源 AF3 类模型,进入设计过滤结构预测来源
33Brian CoventryUniversity of Washington IPDRIFDock / RIFGen 旋转异构体相互作用场对接与曹龙兴 Nature 2022 minibinder;代码 https://github.com/rifdock/rifdock物理计算来源
34Bassil Dahiyat 巴西尔·达希亚特Xencor(创业) / Caltech 博士FSD-1:全自动计算设计的 28 残基蛋白Mayo 实验室;后创办抗体工程公司 Xencor物理计算来源
35Rhiju DasStanford UniversityRNA modeling, RNA design, nucleic acids, structure prediction, biophysics, RNA folding, computational tools.RosettaCommons 实验室名录相邻领域来源
36Justas DauparasUniversity of Washington IPDProteinMPNN / LigandMPNNBaker 实验室;逆折叠把成功率从力场时代抬起生成设计来源
37William DeGrado 威廉·德格拉多UCSFMembrane protein structure, de novo protein design, metalloproteins, chemical biology, neurodegenerative diseases, bacterial sensing, machine learning, drug development.RosettaCommons 实验室名录物理计算来源
38Frank DiMaioUniversity of WashingtonStructure determination, crystal structure refinement, conformational sampling, protein force fields, computational modeling, experimental data analysis.RosettaCommons 实验室名录RosettaCommons来源
39Ken A. DillStony Brook / Laufer CenterHP 格子、可设计性理论折叠理论,不是湿实验设计家前提/评估来源
40Jiayi Dou 窦嘉毅ShanghaiTech UniversityDe novo protein design, neuroscience, light-sensitive proteins, ion channels, membrane fusion, optogenetics.RosettaCommons 实验室名录中国节点来源
41Kevin DrewUniversity of Illinois ChicagoMacromolecular assemblies, protein-nucleic acid interactions, biochemical approaches, computational modeling, proteomics, structural biology, human disease.RosettaCommons 实验室名录RosettaCommons来源
42Roland DunbrackFox Chase Cancer CenterStructural bioinformatics, cryo-EM, X-ray crystallography, deep learning, GTPases, kinases, antibodies, cancer biology, precision medicine.RosettaCommons 实验室名录相邻领域来源
43Richard EvansGoogle DeepMindAlphaFold-Multimer / AF3 复合物Jumper 合作者;复合物预测进入 binder 过滤结构预测来源
44Noelia FerruzCRG / 后迁ProtGPT2 等序列生成开源蛋白 LM 生成生成设计来源
45Eric FischerDana Farber Cancer InstituteUbiquitin proteasome system, structural biology, cell biology, high-throughput biology, targeted protein degradation, SPLiNTs, drug development.RosettaCommons 实验室名录相邻领域来源
46Woody FischerUniversity of Washington IPDRFdiffusion All-Atom 实验Baker 实验室全原子设计生成设计来源
47Sarel Fleishman 萨雷尔·弗莱什曼Weizmann Institute of ScienceComputational protein design, molecular recognition, antibody engineering, enzyme design, membrane protein interactions, experimental characterization.Baker 博士后;魏茨曼独立;Scala BiodesignRosettaCommons来源
48Ora FurmanHebrew UniversityPeptide-protein interactions, computational modeling, structural prediction, molecular interactions, evolutionary analysis, machine learning.RosettaCommons 实验室名录RosettaCommons来源
49Anum GlasgowColumbia UniversityComputational protein design, biophysical techniques, conformational switching, signal-responsive proteins, protein engineering, ligand binding, therapeutic proteins.RosettaCommons 实验室名录RosettaCommons来源
50Cameron GlasscockRice UniversityProtein design, protein-nucleic acid interactions, synthetic biology, artificial intelligence, machine learning, high-throughput biochemistry, functional assays.RosettaCommons 实验室名录RosettaCommons来源
51Haipeng Gong 龚海鹏清华大学蛋白质结构预测与膜蛋白建模清华结构计算;与设计过滤相邻中国节点来源
52Casper GoverdeEPFL Correia labBindCraft 共同作者与 Pacesa / Correia生成设计来源
53Jeffrey J. GrayJohns Hopkins UniversityProtein docking, antibody engineering, glycoproteins, membrane proteins, Rosetta, computational modeling, protein-surface interactions, disease research.RosettaCommons 实验室名录RosettaCommons来源
54Gevorg GrigoryanDartmouth College / Generate BiomedicinesProtein function, structure-based design, protein-protein recognition, allostery, computational modeling, protein engineering.RosettaCommons 实验室名录产业/创业来源
55Dominik GrontUniversity of WarsawStructural bioinformatics, coarse-grained modeling, multiscale modeling, algorithm development, biomolecular dynamics, BioShell software.RosettaCommons 实验室名录RosettaCommons来源
56Alican GulsevinButler CollegeComputational modeling, membrane proteins, protein interactions, structural biology, genomic data, human health.RosettaCommons 实验室名录RosettaCommons来源
57Pehr HarburyStanford University参数化 coiled-coil 骨架与 Peter Kim、Tom Alber 的 packing 几何理性设计来源
58Joe HarrisonUniversity of the PacificUbiquitin interactions, protein inhibitors, ubiquitin pathways, computational modeling, protein design.RosettaCommons 实验室名录RosettaCommons来源
59Demis Hassabis 德米斯·哈萨比斯Google DeepMind / Isomorphic LabsAlphaFold 项目领导;Isomorphic 药物结构与 Baker 同获 2024 诺贝尔化学奖的另一半结构预测来源
60David HausslerUniversity of California, Santa CruzGenomics, bioinformatics, machine learning, statistical methods, genome assembly, UCSC Genome Browser, hidden Markov models, cancer biomarkers.RosettaCommons 实验室名录相邻领域来源
61Jim HavranekDistinguished Scientist, Arzedacomputational protein designRosettaCommons 实验室名录产业/创业来源
62Thomas HayesEvolutionaryScaleESM3 多模态生成esmGFP 实验声称;序列-结构-功能联合生成生成设计来源
63Michael Hecht 迈克尔·赫克特Princeton University二元密码(polar/nonpolar patterning)序列模式而非全原子能量的 de novo 路径理性设计来源
64Brian HieStanford University蛋白 LM 进化 / ESM 应用Meta ESM 合作后独立实验室结构预测来源
65Liang Hong 洪亮上海交通大学Venus 系列:功能预测、AI 定向进化、酶挖掘交大物理/药学交叉;Venus-Pod 数据集与湿实验一体机中国节点来源
66Scott HorowitzUniversity of DenverCitizen science, scientific gaming, Foldit, structural biology, experimental data integration, protein folding, nucleic acids, biophysics.RosettaCommons 实验室名录相邻领域来源
67Parisa HosseinzadehUniversity of OregonComputational tools, protein design, peptide design, molecular modeling, biomedical applications.RosettaCommons 实验室名录RosettaCommons来源
68Chloe HsuUC Berkeley / Meta 实习ESM-IF1 逆折叠Listgarten 轨道;非 Baker 训练生成设计来源
69Possu Huang 黄柏甦Stanford UniversityDe novo protein design, machine learning, molecular motors, protein origami, biosensors, protein engineering, computational modeling, experimental validation.Caltech Mayo 博士;Baker 博士后;后 Stanford 独立实验室物理计算来源
70Thomas HuberAustralian National UniversityGenetic code expansion, paramagnetic NMR, structural bioinformatics, protein design.RosettaCommons 实验室名录RosettaCommons来源
71Birte HöckerUniversity of BayreuthProtein evolution, protein design, small molecule receptors, enzymes, biosensors, molecular machines, artificial motor proteins, small molecule binding.RosettaCommons 实验室名录RosettaCommons来源
72John IngrahamGenerate BiomedicinesChroma 可编程生成Grigoryan / Generate;几何条件生成的产业实现生成设计来源
73Max JaderbergIsomorphic LabsAF 衍生的药物设计平台DeepMind → Isomorphic产业/创业来源
74Robert JeffersonKing's College LondonMembrane protein complexes, quaternary stability, receptor engineering, synthetic biology.RosettaCommons 实验室名录RosettaCommons来源
75Ramesh JhaLos Alamos National LabComputational protein design, synthetic biology, chimeric proteins, biomanufacturing, bioremediation, diagnostics, allosteric biosensors, protein engineering.RosettaCommons 实验室名录RosettaCommons来源
76Lukasz JoachimiakUniversity of Texas SouthwesternProtein folding, amyloid, molecular recognition, chaperones, misfolding, computational methods, structural biology.RosettaCommons 实验室名录RosettaCommons来源
77David JuergensUniversity of Washington IPDRFdiffusion 共同一作Baker 实验室生成模型生成设计来源
78John Jumper 约翰·贾珀Google DeepMindAlphaFold2 结构预测2024 诺贝尔化学奖(预测半部);设计闭环的 in silico 过滤器结构预测来源
79John KaranicolasHead of Computational Drug Discovery, AbbViecomputational protein designRosettaCommons 实验室名录产业/创业来源
80Amy Keating 艾米·基廷MIT螺旋肽/coiled-coil 特异性、计算界面与 Grigoryan 等在识别规则上交叉物理计算来源
81Elizabeth KelloggSt. Jude's Childrens HospitalMolecular organization, nucleic acid interactions, dynamic complexes, Rosetta, cryo-EM, structural biology.RosettaCommons 实验室名录RosettaCommons来源
82Sagar KhareRutgers UniversityComputational biophysics, enzymology, molecular biology, protein design, machine learning, directed evolution, molecular recognition.RosettaCommons 实验室名录RosettaCommons来源
83Firas KhatibUMass DartmouthCitizen science, crowdsourcing, computational biology, public engagement, scientific problem-solving.RosettaCommons 实验室名录相邻领域来源
84Alena KhmelinskaiaLMU Munich, GermanyProtein self-assembly, de novo protein design, biophysical methods, structural flexibility, dynamic protein materials, cargo encapsulation.RosettaCommons 实验室名录RosettaCommons来源
85Neil KingUniversity of WashingtonProtein-based nanomaterials, molecular machines, computational protein design, molecular recognition, structural biology, disease treatment, biophysical characterization.Baker 博士后;Icosavax / SKYCovioneRosettaCommons来源
86Nobuyasu Koga 古贺信康Institute for Molecular ScienceDe novo protein design, monomeric structures, oligomeric structures, novel protein topology.RosettaCommons 实验室名录物理计算来源
87Rie Koga 古贺理惠Institute for Molecular Science规则化骨架/蓝图设计与 Nobuyasu Koga 长期合作;Baker 博士后谱系物理计算来源
88Simon KohlLatent Labs(此前 DeepMind)生成式蛋白质模型创业AlphaFold 作者之一后创办 Latent Labs产业/创业来源
89Pushmeet KohliGoogle DeepMindAlphaFold / AlphaProteo 科学领导DeepMind 科学团队;AlphaProteo 预印本结构预测来源
90Ivan Korendovych 伊万·科连多维奇Syracuse University / 后迁de novo 酶/金属蛋白;与 DeGrado 合写 2020 回顾DeGrado 合作者理性设计来源
91Tanja Kortemme 塔尼亚·科特梅UCSFEngineered biological systems, molecular design, computational biology, cellular function, organismal fitness, predictive modeling, synthetic biology.Baker 博士后;UCSF 独立;Cell 2024 述评物理计算来源
92Huong KratochvilUniversity of North Carolina at Chapel HillMembrane protein design, functional protein design, molecular recognition, immune response.RosettaCommons 实验室名录RosettaCommons来源
93Rohith KrishnaUniversity of Washington IPDRoseTTAFold All-Atom / RFdiffusion-AA全原子把核酸、离子、小分子纳入同一生成器生成设计来源
94Georg KuenzeLeipzig UniversityProtein modeling, membrane proteins, signaling mechanisms, enzyme design, molecular dynamics, molecular signaling, biotechnology, biophysics, drug targeting.RosettaCommons 实验室名录RosettaCommons来源
95Brian Kuhlman 布莱恩·库尔曼UNC Chapel HillProtein design, protein-protein interactions, computational modeling, experimental validation, Rosetta, vaccine design, structural biology.Baker 博士后 1999–2002;Top7 一作物理计算来源
96Dan KulpThe Wistar InstituteProtein engineering, antigen design, antibody interactions, immunogen design, receptor interfaces, vaccine design, antigen stabilization.RosettaCommons 实验室名录产业/创业来源
97Jason LabonteNotre Dame of Maryland UniversityComputational modeling, carbohydrates, post-translational modifications (PTMs), protocol development, data mining, organic chemistry, biochemistry.RosettaCommons 实验室名录RosettaCommons来源
98Luhua Lai 来鲁华北京大学计算药物/蛋白质设计与酶工程中国计算设计长期节点,与 Baker 谱系并行中国节点来源
99Andrew Leaver-FayMolecular Modeling Lead, Genentechcomputational protein designRosettaCommons 实验室名录产业/创业来源
100Gyu Rie Lee 李奎利Korea Advanced Institute of Science and TechnologyProtein design, deep learning, small molecules, epigenetics, post-translational modifications.RosettaCommons 实验室名录RosettaCommons来源
101Sarah LewisMicrosoft ResearchBioEmu 一作与 Noé、Clementi 等结构预测来源
102Jiang LinUCLAProtein aggregation, neurodegenerative diseases, allosteric inhibitors, prion-like transmission, blood-brain barrier, therapeutic design, protein inhibitors.RosettaCommons 实验室名录RosettaCommons来源
103Zeming Lin 林泽明EvolutionaryScale / 此前 MetaESM-2 / ESMFoldMSA-free 结构预测,设计中的快速过滤结构预测来源
104Steffen LindertOhio State UniversityComputational biophysics, protein structure prediction, protein dynamics, protein-ligand interactions, drug discovery, cryo-EM, covalent labeling, ion mobility, mass spectrometry.RosettaCommons 实验室名录RosettaCommons来源
105Sidney LisanzaUniversity of Washington IPDProteinGenerator 序列空间扩散Baker 实验室多状态功能设计生成设计来源
106Haiyan Liu 刘海燕中国科学技术大学SCUBA / ABACUS 骨架与序列设计与陈泉构成独立于 Baker 栈的中国闭环;Nature 2022 SCUBA中国节点来源
107Sen Liu 刘森Hubei University of TechnologyProtein structure, protein regulation, computational modeling, signaling networks, drug development, molecular tools, cellular engineering.RosettaCommons 实验室名录中国节点来源
108Ajasja LjubeticNational Institute of Chemistry (Slovenia)De novo protein design, dynamic protein assemblies, coiled coils, protein robotics, computational modeling, Rosetta software, protein engineering.RosettaCommons 实验室名录RosettaCommons来源
109Nir LondonWeizmann Institute of ScienceComputational protein design, chemical biology, drug discovery, protein-ligand interactions, covalent inhibitors, enzyme targeting.RosettaCommons 实验室名录RosettaCommons来源
110Peilong Lu 卢培龙Westlake University (Hangzhou, China)Transmembrane proteins, computational design, nanopores, ion channels, ligand-binding proteins, oligomerization, high-throughput screening, protein engineering.Baker 博士后 2015–19;西湖独立 PI中国节点来源
111Jianzhu Ma 马剑竹清华大学 AIR抗体/肽生成、PocketXMol博士导师许锦波;Helixon 历史任职中国节点来源
112Ali MadaniProfluentProGen 自回归序列生成Salesforce → Profluent;与结构扩散平行的序列 LM 路径生成设计来源
113Tobias MadlMedical University of GrazIntegrative structural biology, metabolomics, NMR spectroscopy, disordered proteins, transcription factors, RNA-binding proteins, cancer, neurodegeneration, ageing.RosettaCommons 实验室名录相邻领域来源
114Lars MalmstromUniversity of ZurichProteome modeling, computational biology, systems biology, machine learning, statistical analysis, high-performance computing, information management.RosettaCommons 实验室名录相邻领域来源
115Stephen Mayo 斯蒂芬·梅奥CaltechORBIT 死端消除 / 自动化侧链设计Dahiyat 的导师;1997 Science 全自动设计物理计算来源
116Joshua MeierChai Discovery(此前 Meta / Absci)ESM-1;Chai-1/2ESM 共同作者 → Absci → Chai 创始人产业/创业来源
117Jens MeilerVanderbilt University / Leipzig UniversityProtein-ligand interactions, computational modeling, drug design, membrane proteins, structure elucidation, experimental validation, protein structure.RosettaCommons 实验室名录RosettaCommons来源
118Jeremy MillsArizona State UniversityProtein design, fluorescent proteins, non-canonical amino acids, fluorogenic probes, protein-protein interactions, metalloproteins, enzyme engineering.RosettaCommons 实验室名录RosettaCommons来源
119Milot MirditaSeoul National UniversityColabFoldSteinegger / Ovchinnikov 合作结构预测来源
120James MoodyBrigham Young UniversityProtein modeling, enzyme engineering, protein complexes, crystallization chaperones, protein polymers, structural enzymology, protein crystallization.RosettaCommons 实验室名录RosettaCommons来源
121John Moult 约翰·莫尔特University of Maryland / CASPCASP 盲测评估把预测精度变成可比较的公共考试,间接校准设计过滤前提/评估来源
122Vikram MulliganFlatiron Institute非天然氨基酸 / 大环肽 Rosetta与 Renfrew 的异聚物设计相邻物理计算来源
123Christoffer NornBioinnovation Institute DenmarkProtein design, protein binders, minibinders, membrane proteins, high-throughput screening, computational modeling, protein-protein interactions.RosettaCommons 实验室名录产业/创业来源
124Dror NoyMIGAL - Galilee Research InstituteProtein design, photosynthesis, protein-cofactor complexes, computational modeling, minimal protein systems, energy conversion.RosettaCommons 实验室名录RosettaCommons来源
125Frank NoéMicrosoft Research AI for ScienceBioEmu 平衡系综与自由能动力学/热力学模拟的深度学习替代;Science 2025结构预测来源
126Matt O'MearaUniversity of MichiganComputational modeling, molecular simulations, deep neural networks, molecular interactions, chemical modeling, biochemical validation.RosettaCommons 实验室名录RosettaCommons来源
127Gustav OberdorferGraz University of TechnologyDe novo protein design, enzyme design, biomedical applications, biotechnology, computational modeling, protein engineering, structural biology.RosettaCommons 实验室名录RosettaCommons来源
128Byung-Ha OhKAIST KoreaComputational protein design, antibody engineering, protein-protein interactions, protein binders, molecular modeling, therapeutic proteins.RosettaCommons 实验室名录RosettaCommons来源
129Sergey OvchinnikovMITProtein evolution, statistical modeling, phylogenetics, genomics, protein structure prediction, metagenomics, conformational landscapes, protein-protein interactions, multicellularRosettaCommons 实验室名录结构预测来源
130Martin PacesaEPFL Correia labBindCraft 端到端 binderCorreia 实验室;把 AF2 梯度与设计焊成一条管线生成设计来源
131Jesper PallesenIndiana UniversityCryo-electron microscopy (Cryo-EM), cancer, HIV, coronaviruses, SARS-CoV-2, MERS-CoV, filovirus, ebolavirus, structural biology.RosettaCommons 实验室名录相邻领域来源
132Fabio ParmeggianiUniversity of Bristol, UKHybrid computational-experimental methods, protein design, repeat proteins, modular systems, nanomaterials, spatial control, cellular behavior.RosettaCommons 实验室名录RosettaCommons来源
133Brian PierceUniversity of MarylandProtein design, immune recognition, T cell receptors, antibodies, vaccine design, computational docking, Rosetta, ZAFFI, ZDOCK, immunotherapeutics.RosettaCommons 实验室名录RosettaCommons来源
134Nicholas PolizziDana-Farber Cancer InstituteLigand binding, molecular recognition, de novo protein design.RosettaCommons 实验室名录RosettaCommons来源
135Florian PraetoriusInstitute of Science and Technology AustriaProtein design, DNA origami, nucleic acids, deep learning, ProteinMPNN, Rosetta, biomolecular assemblies.RosettaCommons 实验室名录RosettaCommons来源
136Erik ProckoDirector of Discovery, Cyrus Biotechnologycomputational protein designRosettaCommons 实验室名录产业/创业来源
137Vatsan RamanUniversity of Wisconsin-MadisonComputational biology, synthetic biology, protein engineering, high-throughput experiments, single-pot assays, mutational analysis, systems biology.RosettaCommons 实验室名录RosettaCommons来源
138Cesar Ramirez-SarmientoPontificia Universidad Catolica de ChileProtein engineering, computational modeling, metamorphic proteins, PET hydrolysis, enzyme evolution, synthetic biology, oxygen carriers, photosynthetic microorganisms.RosettaCommons 实验室名录RosettaCommons来源
139Barak RavehHebrew UniversityDynamic biological systems, quantitative biology, systems biology, computational modeling, biological interactions, data integration.RosettaCommons 实验室名录RosettaCommons来源
140Lynne Regan 琳恩·里根University of Edinburgh(此前 Yale)四螺旋束 de novo 设计与 DeGrado 合作的早期 uniquely folded 螺旋束理性设计来源
141P. Douglas RenfrewFlatiron Institute / Simons FoundationProtein folding, heteropolymer design, exotic amino acids, peptide macrocycles, enzyme engineering, nanomaterials, computational modeling, quantum computing, deep learning.RosettaCommons 实验室名录RosettaCommons来源
142David RichardsonDuke UniversityMolProbity / all-atom contact与 Jane Richardson 共享实验室前提/评估来源
143Jane Richardson 简·理查森Duke University蛋白质解剖图、ribbon 图、结构质量标准为后续设计提供可检验的结构语言前提/评估来源
144Alexander Rives 亚历山大·里夫斯EvolutionaryScale(此前 Meta FAIR)ESM 蛋白语言模型序列世界模型路径,与结构扩散平行结构预测来源
145Gabe RocklinNorthwestern UniversityHigh-throughput protein design, computational modeling, protein biophysics, protein therapeutics, display selections, mass spectrometry proteomics, next-generation sequencing, protRosettaCommons 实验室名录RosettaCommons来源
146Olaf RonnebergerGoogle DeepMindAlphaFold 架构(Evoformer)计算机视觉转入结构预测结构预测来源
147Rituparna SamantaUniversity of South FloridaMembrane proteins, protein-protein interface, molecular simulation, field-based simulation, protein self-assembly.RosettaCommons 实验室名录RosettaCommons来源
148Deanne SammondSenior Principal Scientist, Pfizercomputational protein designRosettaCommons 实验室名录产业/创业来源
149William SchiefVice President, Protein Design, Modernacomputational protein designRosettaCommons 实验室名录产业/创业来源
150Torben SchiffnerScripps Research InstituteComputational protein design, high-throughput screening, immunogens, epitope grafting, glycan masking, multimerization, vaccine development, pathogens.RosettaCommons 实验室名录RosettaCommons来源
151Clara SchoederLeipzig UniversityComputational protein design, immunotherapeutics, Adeno-associated virus (AAV), chimeric antigen receptors (CARs), single-chain variable fragments (scFvs), vaccine design, epitope RosettaCommons 实验室名录RosettaCommons来源
152Tom SercuEvolutionaryScale / 此前 MetaESM 训练与规模化Rives 核心合作者结构预测来源
153Nik SgourakisChildrens Hospital of Pennsylvania (CHOP)Structural biology, NMR spectroscopy, X-ray crystallography, biophysics, biochemical techniques, computational modeling, protein interactions.RosettaCommons 实验室名录相邻领域来源
154Yunyu Shi 施蕴渝中国科学技术大学计算结构生物学(国内开拓)刘海燕/陈泉 SCUBA 路线的学科前提,本人不以 de novo 设计为标签前提/评估来源
155Michael ShirtsUniversity of Colorado BoulderNanoscale materials, biomimetic materials, foldamers, computational modeling, Rosetta, chemical design, molecular simulations.RosettaCommons 实验室名录RosettaCommons来源
156Justin SiegelUC DavisComputational enzyme design, enzyme catalysis, genetic methods, chemical methods, novel catalysts.RosettaCommons 实验室名录RosettaCommons来源
157Daniel-Adriano SilvaNeoleukin Therapeutics(历史)de novo 细胞因子 mimeticBaker 实验室 → Neoleukin 创业产业/创业来源
158Sam SinaiNabla Bio适应度景观 / 生成设计与 Biswas 共同创办产业/创业来源
159Joanna SluskyUniversity of KansasMembrane proteins, computational protein design, protein characteristics, membrane protein design, protein-membrane interactions.RosettaCommons 实验室名录RosettaCommons来源
160Colin SmithWesleyan UniversityProtein structure, atomic-level mechanisms, second-shell interactions, structural propagation, protein conformational changes, therapeutic protein engineering, structural biology.RosettaCommons 实验室名录RosettaCommons来源
161Yifan Song 宋一凡Cyrus BiotechnologyRosetta 产业化Baker 谱系软件公司产业/创业来源
162Martin SteineggerSeoul National UniversityMMseqs2 / ColabFold / Foldseek把 AF2 变成人人可跑的设计前置结构预测来源
163Lance StewartInstitute for Protein DesignIPD 转化与产业接口Baker 研究所运营/转化产业/创业来源
164Eva-Maria StrauchWashington University of St. LouisProtein engineering, viral proteins, structural design, oligo-synthesis, next-generation sequencing, immunogen design, protein chemistry.RosettaCommons 实验室名录RosettaCommons来源
165Terra SztainUniversity of MichiganComputer-aided molecular design, molecular dynamics, machine learning, drug design.RosettaCommons 实验室名录RosettaCommons来源
166Brian L. TrippeColumbia / 此前 MITRFdiffusion 统计/扩散合作生成模型理论进入实验设计论文生成设计来源
167Kathryn TunyasuvunakoolGoogle DeepMindAlphaFold DB 人类蛋白质组把预测变成公共结构库,供设计检索结构预测来源
168Robert VerkuilEvolutionaryScale / 此前 MetaESMFold 推理与表示Lin/Rives 合作结构预测来源
169Anastassia VorobievaVIB-VUB BelgiumMembrane proteins, transmembrane beta-barrels, computational protein design, nanopore engineering, protein stability, single-molecule sensing, beta-barrel vaccines.RosettaCommons 实验室名录RosettaCommons来源
170Bjorn WallnerLinkoping University SwedenBioinformatics, molecular energy functions, protein-protein interactions, molecular sampling, machine learning.RosettaCommons 实验室名录RosettaCommons来源
171Chu Wang 王初Peking UniversityProtein docking, metal-binding proteins, chemoproteomics, bioinformatics, metabolic regulation, enzyme function, computational biology.RosettaCommons 实验室名录中国节点来源
172Jue Wang 王珏University of Washington IPD幻觉设计 / 深度学习骨架Baker 实验室;与 ProteinMPNN 同期 Science生成设计来源
173Joseph L. WatsonUniversity of Washington IPDRFdiffusion 一作与 Juergens、Bennett、Yim、Trippe 共同把 RF 变成扩散生成器生成设计来源
174Hannah Wayment-SteeleUniversity of Wisconsin-MadisonBiomolecular dynamics, evolution, deep learning, data science, NMR spectroscopy, millisecond-scale motions.RosettaCommons 实验室名录RosettaCommons来源
175Tim WhiteheadUniversity of Colorado BoulderProtein engineering, antibodies, molecular recognition, next-generation sequencing, enzyme specificity, plant synthetic biology, renewable biomass.RosettaCommons 实验室名录RosettaCommons来源
176Basile WickyUniversity of Washington IPD / 后迁深度学习组装体的实验确认与 Dauparas 同期 Science 实验闭环生成设计来源
177Jeremy WohlwendMITBoltz 开源结构预测Corso / MIT 几何 ML结构预测来源
178Dek Woolfson 德克·伍尔夫森University of Bristolcoiled-coil 规则、CC+ 数据库、理性组装与 Baker/DeGrado 平行的螺旋组装传统;2021 JMB 简史理性设计来源
179Wenzhi Xiao 肖文智ByteDance Seed / ProtenixProtenix;PXDesign binder开源 AF3 类模型 + 公开 binder 湿实验命中率(预印本)中国节点来源
180Chunfu Xu 徐春福National Institute of Biological SciencesProtein design, protein-nucleic acid interactions, synthetic biology, artificial intelligence, machine learning, high-throughput biochemistry, functional assays.RosettaCommons 实验室名录中国节点来源
181Jinbo Xu 许锦波分子之心 / 此前 Toyota Technological InstituteRaptorX、深度学习接触图AF2 之前的深度结构预测主线之一结构预测来源
182Jianyi Yang 杨建溢Shandong UniversityStructure prediction, cryo-EM, artificial intelligence, trRosetta, trRosettaRNA, I-TASSER.RosettaCommons 实验室名录中国节点来源
183Vladimir Yarov-YarovoyUC DavisSodium channels, neuromodulation, epilepsy, chronic pain, drug design, electrophysiology, structural biology.RosettaCommons 实验室名录RosettaCommons来源
184Andy YehUniversity of California, Santa CruzLuciferase, biosensor, enzyme design, luminescence imaging, molecular imaging, clinical diagnostics, drug discovery.RosettaCommons 实验室名录RosettaCommons来源
185Joe YesselmanUniversity of Nebraska LincolnRNA structure, RNA design, nanomachines, RNA thermodynamics, predictive modeling, RNA energetics.RosettaCommons 实验室名录相邻领域来源
186Jason YimMIT / 后迁SE(3) 扩散、RFdiffusion 合作与 Baker 实验室交叉的几何生成生成设计来源
187Fajie Yuan 原发杰西湖大学SaProt / ProTrek / 文本条件蛋白设计西湖表示学习,与曹/卢实验组地理相邻中国节点来源
188Vinicius ZambaldiGoogle DeepMindAlphaProteo binder 生成(预印本)DeepMind 设计侧;代码受限,与开源 RFdiffusion 对照生成设计来源
189Alexandre ZanghelliniArzeda计算酶设计产业化Baker 酶设计谱系 → Arzeda;Havranek 现任职产业/创业来源
190Jianyang Zeng 曾建洋西湖大学 / HelixonAI 药物与结构预测创业与西湖设计集群地理相邻的产业节点中国节点来源
191Haicang Zhang 张海仓中科院计算技术研究所CarbonDesign Inverseformer 逆折叠与卜东波共同通讯;Nat Mach Intell 2024中国节点来源
192Yang Zhang 张阳National University of Singapore(此前 UM)I-TASSER / 结构预测与设计交叉与 Jianyi Yang 的 I-TASSER/trRosetta 谱系结构预测来源

显示 192 / 192

10. 2020 年后的设计闭环

IPD 软件页把当代栈写成三件套:RoseTTAFold 预测、RFdiffusion 出骨架、ProteinMPNN 填序列。实验论文反复使用同一句:先生成,再 MPNN,再用 AF2 的 pAE / pLDDT / RMSD 丢掉自洽失败的候选。Dauparas 等报告 ProteinMPNN 在天然骨架上 52.4% 序列恢复,对照 Rosetta 32.9%,并用晶体、电镜和功能实验救活单体、环状寡聚体、四面体纳米粒和 binder。Watson 等报告 RFdiffusion 在多种条件下优于先前方法,并给出计算直接得到的皮摩尔级 binder(取决于靶标,不能写成普遍成功率)。

Generate 的 Chroma 把生成写成可编程几何指令。EvolutionaryScale 的 ESM3 走序列、结构、功能多模态,esmGFP 是其实验旗舰;证据强度取决于该文的湿实验部分,不取决于演示视频。DeepMind 的 AlphaProteo 仍是 2024 预印本且代码受限,实验公开密度低于 RFdiffusion。BindCraft 把生成和预测合成一条可反传的 binder 管线,作者单位在 EPFL Correia 组。功能化不独占西雅图。

西湖把同一套 IPD 软件用在不同功能上:卢培龙组的跨膜与毒素中和迷你蛋白,曹龙兴组的开关与 2026 年小分子响应寡聚体,陈子博组的蛋白电路。实验在国内完成,方法引用仍指向 IPD 仓库。代码边比师生边扩散得快。

环节代表方法关键人物已有实验证据常见误读
骨架生成RFdiffusion / RFAA / BindCraft / ChromaWatson, Juergens, Bennett, Krishna, Pacesa, IngrahamNature/Science 结构与结合把无条件生成当功能设计
刚体对接 / RIFRIFGen + RIFDock;RosettaDockCoventry, Cao, GlasscockNature 2022 minibinder;2025 DNA 结合管线仍用写成已被扩散完全替代
逆折叠ProteinMPNN / LigandMPNN / SolubleMPNNDauparas, Anishchenko, BakerScience 2022;Nat Methods 2025;IPD 称 >100 配体相关实验序列恢复 ≠ 表达溶解
过滤AF2 / RF / ESMFold / AF3 / BoltzJumper, Baek, Lin, Abramson, Corso预测精度有 CASP/论文;作为设计过滤器是相关而非因果把预测作者写成设计家
功能湿实验表达、结合、晶体、细胞、动物Baker, King, Cao, Lu, Correia, Fleishman, Huang因靶而异;疫苗与 minibinder 多于酶用 in silico 命中率代替体内

11. 综合讨论

方法史与人事枢纽可以对上。没有 DeGrado 的可合成对象,Mayo 的组合优化没有实验落点。没有 Kuhlman/Baker 的 Top7,蓝图规则缺少新拓扑的存在性证明。没有 AF2,MPNN 的筛选比到不了现在的实验室吞吐。没有 Commons 把博士后变成 PI,西湖和洛桑不会使用同一套输入文件。中国北方轴给出反例:SCUBA 与 CarbonDesign、TopoDiff 可以在不经过 Baker 人事的情况下发表晶体或逆折叠基准。

诺贝尔奖把设计与预测分开授奖。设计论文把预测器用成过滤器:pLDDT 和 pAE 高,不等于写出有功能的蛋白。Generate、Isomorphic、Profluent、Nabla、Cyrus、Arzeda 使用同一套开源权重,发表密度低于 IPD,外部不能按 Watson 2023 的标准比较湿实验成功率。Helixon、Protenix、分子之心是公司名,不能替代湿实验第一作者。动力学模型给出构象分布与 ΔG,还不能替代结合与催化测定。

12. 未解决问题

13. 结论

RQ1。 分期按可检验产物:理性螺旋束,全原子新拓扑(Top7),可筛选功能蛋白。2020 年后的变量是过滤器吞吐量。热力学能量函数(Rosetta、SCUBA)与平衡系综(BioEmu)仍服务稳态;动力学接触图与开关设计正在进入目标函数,尚未成为验收标准。

RQ2。 192 人之间的边是稀疏的。Baker/IPD、DeGrado、Mayo、Correia、DeepMind/ESM 通过博士后、公共代码和创业相连。中国有两条组织:交大–中科大–计算所–清华–北大的数据驱动与功能工程,西湖的 Baker 博士后实验闭环。师生边未公开则保持未证实。

RQ3。 AF2/RF/ESM 改写候选过滤和部分梯度设计(BindCraft),不改写物理原则。有晶体或结合数据的主张集中在 RFdiffusion、ProteinMPNN、SCUBA、TopoDiff 的实验子集,以及西湖功能蛋白。AlphaProteo 与部分全原子酶主张仍为预印本或有限实验。预测作者与设计作者的职业轨迹大部分不相交,与 2024 年诺贝尔奖的授奖结构一致。

参考文献

  1. Dahiyat BI, Mayo SL. "De novo protein design: fully automated sequence selection." Science, 1997.
  2. Kuhlman B, Dantas G, Ireton GC, Varani G, Stoddard BL, Baker D. "Design of a novel globular protein fold with atomic-level accuracy." Science, 2003.
  3. Huang PS, Boyken SE, Baker D. "The coming of age of de novo protein design." Nature, 2016.
  4. Korendovych IV, DeGrado WF. "De novo protein design, a retrospective." Q Rev Biophys, 2020.
  5. Woolfson DN. "A brief history of de novo protein design: minimal, rational, and computational." J Mol Biol, 2021.
  6. Jumper J, et al. "Highly accurate protein structure prediction with AlphaFold." Nature, 2021.
  7. Baek M, et al. "Accurate prediction of protein structures and interactions using a three-track neural network." Science, 2021.
  8. Dauparas J, et al. "Robust deep learning–based protein sequence design using ProteinMPNN." Science, 2022.
  9. Lin Z, et al. "Evolutionary-scale prediction of atomic-level protein structure with a language model." Science, 2023.
  10. Watson JL, Juergens D, Bennett NR, et al. "De novo design of protein structure and function with RFdiffusion." Nature, 2023.
  11. Kortemme T. "De novo protein design—From new structures to programmable functions." Cell, 2024.
  12. Chu AE, Lu T, Huang PS. "Sparks of function by de novo protein design." Nat Biotechnol, 2024.
  13. Abramson J, et al. "Accurate structure prediction of biomolecular interactions with AlphaFold 3." Nature, 2024.
  14. Krishna R, Wang Y, Ahern W, et al. "Generalized biomolecular modeling and design with RoseTTAFold All-Atom." Science, 2024.
  15. The Nobel Prize in Chemistry 2024. NobelPrize.org.
  16. Dauparas J, et al. "Atomic context-conditioned protein sequence design using LigandMPNN." Nat Methods, 2025.
  17. Pacesa M, et al. "BindCraft." Nature, 2025. doi:10.1038/s41586-025-09429-6.
  18. Zambaldi V, et al. "AlphaProteo." arXiv:2409.08022, 2024. 预印本.
  19. RosettaCommons. "Labs." https://rosettacommons.org/about/labs/ 抓取于 2026-09-14.
  20. Institute for Protein Design. "Software." https://www.ipd.uw.edu/software/
  21. Baker D. Nobel Prize lecture "De Novo Protein Design," 2024.
  22. Regan L, DeGrado WF. "Characterization of a helical protein designed from first principles." Science, 1988.
  23. Kamtekar S, Schiffer JM, Xiong H, Babik JM, Hecht MH. "Protein design by binary patterning of polar and nonpolar amino acids." Science, 1993.
  24. Mirdita M, et al. "ColabFold: making protein folding accessible to all." Nat Methods, 2022.
  25. Huang B, et al. "A backbone-centred energy function of neural networks for protein design." Nature, 2022. (SCUBA)
  26. Liu Y, Wang S, et al. "De novo protein design with a denoising diffusion network independent of pretrained structure prediction models." Nat Methods, 2024. (SCUBA-D)
  27. Ren M, Yu C, Bu D, Zhang H. "Accurate and robust protein sequence design with CarbonDesign." Nat Mach Intell, 2024.
  28. Zhang Y, Liu Y, Ma Z, Li M, Xu C, Gong H. "Improving diffusion-based protein backbone generation with global-geometry-aware latent encoding." Nat Mach Intell, 2025. (TopoDiff)
  29. Lewis S, Hempel T, Jiménez-Luna J, et al. "Scalable emulation of protein equilibrium ensembles with generative deep learning." Science, 2025. (BioEmu)
  30. Kleiman DE, et al. "ESMDynamic: Fast and accurate prediction of protein conformational dynamics." Nat Commun, 2026.
  31. Cao L, Coventry B, et al. "Design of protein-binding proteins from the target structure alone." Nature, 2022. (RIFDock)
  32. Glasscock CJ, et al. "Computational design of sequence-specific DNA-binding proteins." 公开管线仍以 RIFgen/RIFdock 为第一步,2025.