AI3SD Video: How good are protein structure prediction methods at predicting folding pathways?
AI3SD Video: How good are protein structure prediction methods at predicting folding pathways?
Deep learning has achieved unprecedented success in predicting a protein’s crystal structure, but whether this achievement relates to a better modelling of the folding process is an open question. In this work, we compare the dynamic pathways from six state-of-the-art protein structure prediction methods to experimental folding data. We find evidence of a weak correlation between simulated dynamics and formal kinetics; however, many of the structures of the predicted intermediates are incompatible with available hydrogen-deuterium exchange experiments. These results suggest that recent advances in protein structure prediction do not provide an enhanced understanding of the principles underpinning protein folding.
AI, AI3SD, AI3SD Event, Artificial Intelligence, Machine Intelligence, Machine Learning, ML, Proteins
Rubiera, Carlos Outeiral
fddf7e89-01b0-49cd-93f6-b029b2d0e05c
Kanza, Samantha
b73bcf34-3ff8-4691-bd09-aa657dcff420
Frey, Jeremy G.
ba60c559-c4af-44f1-87e6-ce69819bf23f
Niranjan, Mahesan
5cbaeea8-7288-4b55-a89c-c43d212ddd4f
16 June 2021
Rubiera, Carlos Outeiral
fddf7e89-01b0-49cd-93f6-b029b2d0e05c
Kanza, Samantha
b73bcf34-3ff8-4691-bd09-aa657dcff420
Frey, Jeremy G.
ba60c559-c4af-44f1-87e6-ce69819bf23f
Niranjan, Mahesan
5cbaeea8-7288-4b55-a89c-c43d212ddd4f
Rubiera, Carlos Outeiral
(2021)
AI3SD Video: How good are protein structure prediction methods at predicting folding pathways?
Kanza, Samantha, Frey, Jeremy G. and Niranjan, Mahesan
(eds.)
AI 4 Proteins Seminar Series 2021.
14 Apr - 17 Jun 2021.
(doi:10.5258/SOTON/P0110).
Record type:
Conference or Workshop Item
(Other)
Abstract
Deep learning has achieved unprecedented success in predicting a protein’s crystal structure, but whether this achievement relates to a better modelling of the folding process is an open question. In this work, we compare the dynamic pathways from six state-of-the-art protein structure prediction methods to experimental folding data. We find evidence of a weak correlation between simulated dynamics and formal kinetics; however, many of the structures of the predicted intermediates are incompatible with available hydrogen-deuterium exchange experiments. These results suggest that recent advances in protein structure prediction do not provide an enhanced understanding of the principles underpinning protein folding.
Video
AI4Proteins-Seminar-Series-CarlosOuteiral-160621
- Version of Record
More information
Published date: 16 June 2021
Additional Information:
Carlos Outeiral was born and raised in northern Spain, where he completed his baccalaureate (achieving a National Award) and earned a BSc in Chemistry at the University of Oviedo (achieving the honours of valedictorian and extraordinary award). Following internships at the Technical University of Munich (Germany) and Harvard University (US), he completed a MPhil in Chemistry at the University of Manchester (UK). He is currently a final-year PhD candidate at the University of Oxford (UK), where his research studies novel algorithms to simulate protein folding at scale. Some of his work has examined the prospects of quantum computing in computational biology, and developed pipelines for biologically-inspired protein structure prediction. In his free time, Carlos is passionate about deep tech, entrepreneurship and open source software.
Venue - Dates:
AI 4 Proteins Seminar Series 2021, 2021-04-14 - 2021-06-17
Keywords:
AI, AI3SD, AI3SD Event, Artificial Intelligence, Machine Intelligence, Machine Learning, ML, Proteins
Identifiers
Local EPrints ID: 450196
URI: http://eprints.soton.ac.uk/id/eprint/450196
PURE UUID: e3b8ee0e-5e7c-4ade-b283-d4dfe94f0342
Catalogue record
Date deposited: 15 Jul 2021 16:39
Last modified: 17 Mar 2024 03:51
Export record
Altmetrics
Contributors
Author:
Carlos Outeiral Rubiera
Editor:
Mahesan Niranjan
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics