Learning the shape of protein microenvironments with a holographic convolutional neural network.

Autor:	Pun MN; Department of Physics, University of Washington, Seattle, WA 98195.; The Department for Statistical Physics of Evolving Systems, Max Planck Institute for Dynamics and Self-Organization, Göttingen 37077, Germany., Ivanov A; Department of Physics, University of Washington, Seattle, WA 98195., Bellamy Q; Department of Physics, University of Washington, Seattle, WA 98195., Montague Z; Department of Physics, University of Washington, Seattle, WA 98195.; The Department for Statistical Physics of Evolving Systems, Max Planck Institute for Dynamics and Self-Organization, Göttingen 37077, Germany., LaMont C; The Department for Statistical Physics of Evolving Systems, Max Planck Institute for Dynamics and Self-Organization, Göttingen 37077, Germany., Bradley P; Fred Hutchinson Cancer Center, Seattle, WA 98102.; Department of Biochemistry, University of Washington, Seattle, WA 98195.; Institute for Protein Design, University of Washington, Seattle, WA 98195., Otwinowski J; The Department for Statistical Physics of Evolving Systems, Max Planck Institute for Dynamics and Self-Organization, Göttingen 37077, Germany.; Dyno Therapeutics, Watertown, MA 02472., Nourmohammad A; Department of Physics, University of Washington, Seattle, WA 98195.; The Department for Statistical Physics of Evolving Systems, Max Planck Institute for Dynamics and Self-Organization, Göttingen 37077, Germany.; Fred Hutchinson Cancer Center, Seattle, WA 98102.; Department of Applied Mathematics, University of Washington, Seattle, WA 98105.; Paul G. Allen School of Computer Science and Engineering, University of Washington, Seattle, WA 98195.
Jazyk:	angličtina
Zdroj:	Proceedings of the National Academy of Sciences of the United States of America [Proc Natl Acad Sci U S A] 2024 Feb 06; Vol. 121 (6), pp. e2300838121. Date of Electronic Publication: 2024 Feb 01.
DOI:	10.1073/pnas.2300838121
Abstrakt:	Proteins play a central role in biology from immune recognition to brain activity. While major advances in machine learning have improved our ability to predict protein structure from sequence, determining protein function from its sequence or structure remains a major challenge. Here, we introduce holographic convolutional neural network (H-CNN) for proteins, which is a physically motivated machine learning approach to model amino acid preferences in protein structures. H-CNN reflects physical interactions in a protein structure and recapitulates the functional information stored in evolutionary data. H-CNN accurately predicts the impact of mutations on protein stability and binding of protein complexes. Our interpretable computational model for protein structure-function maps could guide design of novel proteins with desired function. Competing Interests: Competing interests statement:J.O. is employed by Dyno Therapeutics.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu