TY - JOUR
T1 - Accurate prediction of protein structures and interactions using a three-track neural network
AU - Baek, Minkyung
AU - DiMaio, Frank
AU - Anishchenko, Ivan
AU - Dauparas, Justas
AU - Ovchinnikov, Sergey
AU - Lee, Gyu Rie
AU - Wang, Jue
AU - Cong, Qian
AU - Kinch, Lisa N.
AU - Dustin Schaeffer, R.
AU - Millán, Claudia
AU - Park, Hahnbeom
AU - Adams, Carson
AU - Glassman, Caleb R.
AU - DeGiovanni, Andy
AU - Pereira, Jose H.
AU - Rodrigues, Andria V.
AU - Van Dijk, Alberdina A.
AU - Ebrecht, Ana C.
AU - Opperman, Diederik J.
AU - Sagmeister, Theo
AU - Buhlheller, Christoph
AU - Pavkov-Keller, Tea
AU - Rathinaswamy, Manoj K.
AU - Dalwadi, Udit
AU - Yip, Calvin K.
AU - Burke, John E.
AU - Christopher Garcia, K.
AU - Grishin, Nick V.
AU - Adams, Paul D.
AU - Read, Randy J.
AU - Baker, David
N1 - Publisher Copyright:
© 2021 American Association for the Advancement of Science. All rights reserved.
PY - 2021/8/20
Y1 - 2021/8/20
N2 - DeepMind presented notably accurate predictions at the recent 14th Critical Assessment of Structure Prediction (CASP14) conference. We explored network architectures that incorporate related ideas and obtained the best performance with a three-track network in which information at the one-dimensional (1D) sequence level, the 2D distance map level, and the 3D coordinate level is successively transformed and integrated. The three-track network produces structure predictions with accuracies approaching those of DeepMind in CASP14, enables the rapid solution of challenging x-ray crystallography and cryo-electron microscopy structure modeling problems, and provides insights into the functions of proteins of currently unknown structure. The network also enables rapid generation of accurate protein-protein complex models from sequence information alone, short-circuiting traditional approaches that require modeling of individual subunits followed by docking. We make the method available to the scientific community to speed biological research.
AB - DeepMind presented notably accurate predictions at the recent 14th Critical Assessment of Structure Prediction (CASP14) conference. We explored network architectures that incorporate related ideas and obtained the best performance with a three-track network in which information at the one-dimensional (1D) sequence level, the 2D distance map level, and the 3D coordinate level is successively transformed and integrated. The three-track network produces structure predictions with accuracies approaching those of DeepMind in CASP14, enables the rapid solution of challenging x-ray crystallography and cryo-electron microscopy structure modeling problems, and provides insights into the functions of proteins of currently unknown structure. The network also enables rapid generation of accurate protein-protein complex models from sequence information alone, short-circuiting traditional approaches that require modeling of individual subunits followed by docking. We make the method available to the scientific community to speed biological research.
UR - http://www.scopus.com/inward/record.url?scp=85111524391&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85111524391&partnerID=8YFLogxK
U2 - 10.1126/science.abj8754
DO - 10.1126/science.abj8754
M3 - Article
C2 - 34282049
AN - SCOPUS:85111524391
SN - 0036-8075
VL - 373
SP - 871
EP - 876
JO - Science
JF - Science
IS - 6557
ER -