Cargando…

Protein structure prediction by AlphaFold2: are attention and symmetries all you need?

The functions of most proteins result from their 3D structures, but determining their structures experimentally remains a challenge, despite steady advances in crystallography, NMR and single-particle cryoEM. Computationally predicting the structure of a protein from its primary sequence has long be...

Descripción completa

Detalles Bibliográficos
Autores principales: Bouatta, Nazim, Sorger, Peter, AlQuraishi, Mohammed
Formato: Online Artículo Texto
Lenguaje:English
Publicado: International Union of Crystallography 2021
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8329862/
https://www.ncbi.nlm.nih.gov/pubmed/34342271
http://dx.doi.org/10.1107/S2059798321007531
_version_ 1783732582262767616
author Bouatta, Nazim
Sorger, Peter
AlQuraishi, Mohammed
author_facet Bouatta, Nazim
Sorger, Peter
AlQuraishi, Mohammed
author_sort Bouatta, Nazim
collection PubMed
description The functions of most proteins result from their 3D structures, but determining their structures experimentally remains a challenge, despite steady advances in crystallography, NMR and single-particle cryoEM. Computationally predicting the structure of a protein from its primary sequence has long been a grand challenge in bioinformatics, intimately connected with understanding protein chemistry and dynamics. Recent advances in deep learning, combined with the availability of genomic data for inferring co-evolutionary patterns, provide a new approach to protein structure prediction that is complementary to longstanding physics-based approaches. The outstanding performance of AlphaFold2 in the recent Critical Assessment of protein Structure Prediction (CASP14) experiment demonstrates the remarkable power of deep learning in structure prediction. In this perspective, we focus on the key features of AlphaFold2, including its use of (i) attention mechanisms and Transformers to capture long-range dependencies, (ii) symmetry principles to facilitate reasoning over protein structures in three dimensions and (iii) end-to-end differentiability as a unifying framework for learning from protein data. The rules of protein folding are ultimately encoded in the physical principles that underpin it; to conclude, the implications of having a powerful computational model for structure prediction that does not explicitly rely on those principles are discussed.
format Online
Article
Text
id pubmed-8329862
institution National Center for Biotechnology Information
language English
publishDate 2021
publisher International Union of Crystallography
record_format MEDLINE/PubMed
spelling pubmed-83298622021-08-19 Protein structure prediction by AlphaFold2: are attention and symmetries all you need? Bouatta, Nazim Sorger, Peter AlQuraishi, Mohammed Acta Crystallogr D Struct Biol Feature Articles The functions of most proteins result from their 3D structures, but determining their structures experimentally remains a challenge, despite steady advances in crystallography, NMR and single-particle cryoEM. Computationally predicting the structure of a protein from its primary sequence has long been a grand challenge in bioinformatics, intimately connected with understanding protein chemistry and dynamics. Recent advances in deep learning, combined with the availability of genomic data for inferring co-evolutionary patterns, provide a new approach to protein structure prediction that is complementary to longstanding physics-based approaches. The outstanding performance of AlphaFold2 in the recent Critical Assessment of protein Structure Prediction (CASP14) experiment demonstrates the remarkable power of deep learning in structure prediction. In this perspective, we focus on the key features of AlphaFold2, including its use of (i) attention mechanisms and Transformers to capture long-range dependencies, (ii) symmetry principles to facilitate reasoning over protein structures in three dimensions and (iii) end-to-end differentiability as a unifying framework for learning from protein data. The rules of protein folding are ultimately encoded in the physical principles that underpin it; to conclude, the implications of having a powerful computational model for structure prediction that does not explicitly rely on those principles are discussed. International Union of Crystallography 2021-07-29 /pmc/articles/PMC8329862/ /pubmed/34342271 http://dx.doi.org/10.1107/S2059798321007531 Text en © Nazim Bouatta et al. 2021 https://creativecommons.org/licenses/by/4.0/This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.
spellingShingle Feature Articles
Bouatta, Nazim
Sorger, Peter
AlQuraishi, Mohammed
Protein structure prediction by AlphaFold2: are attention and symmetries all you need?
title Protein structure prediction by AlphaFold2: are attention and symmetries all you need?
title_full Protein structure prediction by AlphaFold2: are attention and symmetries all you need?
title_fullStr Protein structure prediction by AlphaFold2: are attention and symmetries all you need?
title_full_unstemmed Protein structure prediction by AlphaFold2: are attention and symmetries all you need?
title_short Protein structure prediction by AlphaFold2: are attention and symmetries all you need?
title_sort protein structure prediction by alphafold2: are attention and symmetries all you need?
topic Feature Articles
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8329862/
https://www.ncbi.nlm.nih.gov/pubmed/34342271
http://dx.doi.org/10.1107/S2059798321007531
work_keys_str_mv AT bouattanazim proteinstructurepredictionbyalphafold2areattentionandsymmetriesallyouneed
AT sorgerpeter proteinstructurepredictionbyalphafold2areattentionandsymmetriesallyouneed
AT alquraishimohammed proteinstructurepredictionbyalphafold2areattentionandsymmetriesallyouneed