Premium
A novel approach to prediction of the 3‐dimensional structures of protein backbones by neural networks
Author(s) -
Bohr H.,
Bohr J.,
Brunak S.,
J. Cotterill R.M,
Fredholm H.,
Lautrup B.,
Petersen S.B.
Publication year - 1990
Publication title -
febs letters
Language(s) - English
Resource type - Book series
SCImago Journal Rank - 1.593
H-Index - 257
eISSN - 1873-3468
pISSN - 0014-5793
ISBN - 1-55860-184-8
DOI - 10.1016/0014-5793(90)80632-s
Subject(s) - artificial neural network , computer science , artificial intelligence , computational biology , chemistry , biology
Three‐dimensional structures of protein backbones have been predicted using neural networks. A feed forward neural network was trained on a class of functionally, but not structurally, homologous proteins, using backpropagation learning. The network generated tertiary structure information in the form of binary distance constraints for the C α atoms in the protein backbone. The binary distance between two C α atoms was 0 if the distance between them was less than a certain threshold distance, and 1 otherwise. The distance constraints predicted by the trained neural network were utilized to generate a folded conformation of the protein backbone, using a steepest descent minimization approach.