Top-down design of protein architectures with reinforcement learning
成果类型：
Article
署名作者：
Lutz, Isaac D.; Wang, Shunzhi; Norn, Christoffer; Courbet, Alexis; Borst, Andrew J.; Zhao, Yan Ting; Dosey, Annie; Cao, Longxing; Xu, Jinwei; Leaf, Elizabeth M.; Treichel, Catherine; Litvicov, Patrisia; Li, Zhe; Goodson, Alexander D.; Rivera-Sanchez, Paula; Bratovianu, Ana -Maria; Baek, Minkyung; King, Neil P.; Ruohola-Baker, Hannele; Baker, David
署名单位：
University of Washington; University of Washington Seattle; University of Washington; University of Washington Seattle; University of Washington; University of Washington Seattle; University of Washington; University of Washington Seattle; Howard Hughes Medical Institute; University of Washington; University of Washington Seattle; University of Washington; University of Washington Seattle; Westlake University; Seoul National University (SNU)
刊物名称：
SCIENCE
ISSN/ISSBN：
0036-8842
DOI：
10.1126/science.adf6591
发表日期：
2023-04-21
页码：
266-273
关键词：
computational design
accurate design
angiopoietin-1
symmetry
antibody
game
cage
bond
go
摘要：
As a result of evolutionary selection, the subunits of naturally occurring protein assemblies often fit together with substantial shape complementarity to generate architectures optimal for function in a manner not achievable by current design approaches. We describe a top-down reinforcement learning- based design approach that solves this problem using Monte Carlo tree search to sample protein conformers in the context of an overall architecture and specified functional constraints. Cryo-electron microscopy structures of the designed disk-shaped nanopores and ultracompact icosahedra are very close to the computational models. The icosohedra enable very-high-density display of immunogens and signaling molecules, which potentiates vaccine response and angiogenesis induction. Our approach enables the top-down design of complex protein nanomaterials with desired system properties and demonstrates the power of reinforcement learning in protein design.