Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction

doi:10.1016/j.sbi.2024.102973

Review

. 2025 Feb:90:102973.

doi: 10.1016/j.sbi.2024.102973. Epub 2025 Jan 4.

Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction

Devlina Chakravarty ¹, Myeongsang Lee ¹, Lauren L Porter ²

Affiliations

¹ National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD, 20894, USA.
² National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD, 20894, USA; Biochemistry and Biophysics Center, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, 20892, USA. Electronic address: porterll@nih.gov.

PMID: 39756261
PMCID: PMC11791787 (available on 2026年02月01日)
DOI: 10.1016/j.sbi.2024.102973

Review

Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction

Devlina Chakravarty et al. Curr Opin Struct Biol. 2025 Feb.

. 2025 Feb:90:102973.

doi: 10.1016/j.sbi.2024.102973. Epub 2025 Jan 4.

Authors

Devlina Chakravarty ¹, Myeongsang Lee ¹, Lauren L Porter ²

Affiliations

¹ National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD, 20894, USA.
² National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD, 20894, USA; Biochemistry and Biophysics Center, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, 20892, USA. Electronic address: porterll@nih.gov.

PMID: 39756261
PMCID: PMC11791787 (available on 2026年02月01日)
DOI: 10.1016/j.sbi.2024.102973

Abstract

In recent years, advances in artificial intelligence (AI) have transformed structural biology, particularly protein structure prediction. Though AI-based methods, such as AlphaFold (AF), often predict single conformations of proteins with high accuracy and confidence, predictions of alternative folds are often inaccurate, low-confidence, or simply not predicted at all. Here, we review three blind spots that alternative conformations reveal about AF-based protein structure prediction. First, proteins that assume conformations distinct from their training-set homologs can be mispredicted. Second, AF overrelies on its training set to predict alternative conformations. Third, degeneracies in pairwise representations can lead to high-confidence predictions inconsistent with experiment. These weaknesses suggest approaches to predict alternative folds more reliably.

Keywords: Alternative conformations; Deep learning; Fold-switching proteins; Machine learning; Metamorphic proteins; Protein structure prediction.

Published by Elsevier Ltd.

PubMed Disclaimer

Conflict of interest statement

Declaration of competing interest The authors declare no conflict of interest.

Update of

Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction.
Chakravarty D, Lee M, Porter LL. Chakravarty D, et al. ArXiv [Preprint]. 2024 Oct 18:arXiv:2410.14898v1. ArXiv. 2024. Update in: Curr Opin Struct Biol. 2025 Feb;90:102973. doi: 10.1016/j.sbi.2024.102973. PMID: 39801626 Free PMC article. Updated. Preprint.

References

1. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A, et al., Highly accurate protein structure prediction with AlphaFold, Nature 596 (2021) 583–589. 10.1038/s41586-021-03819-2. - DOI - PMC - PubMed
1. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A, et al., Applying and improving AlphaFold at CASP14, Proteins 89 (2021) 1711–1721. 10.1002/prot.26257. - DOI - PMC - PubMed
1. Senior AW, Evans R, Jumper J, Kirkpatrick J, Sifre L, Green T, Qin C, Zidek A, Nelson AWR, Bridgland A, et al., Improved protein structure prediction using potentials from deep learning, Nature 577 (2020) 706–710. 10.1038/s41586-019-1923-7. - DOI - PubMed
1. Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, Ronneberger O, Willmore L, Ballard AJ, Bambrick J, et al., Accurate structure prediction of biomolecular interactions with AlphaFold 3, Nature (2024) 10.1038/s41586-024-07487-w. - DOI - PMC - PubMed
1. Ahdritz G, Bouatta N, Floristean C, Kadyan S, Xia Q, Gerecke W, O’Donnell TJ, Berenberg D, Fisk I, Zanichelli N, et al., OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization, Nat Methods (2024) 10.1038/s41592-024-02272-z. - DOI - PMC - PubMed

Publication types

Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions

Substances

Actions

Grants and funding

ZIA LM202011/ImNIH/Intramural NIH HHS/United States

LinkOut - more resources

Full Text Sources
- Elsevier Science

[1] Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A, et al., Highly accurate protein structure prediction with AlphaFold, Nature 596 (2021) 583–589. 10.1038/s41586-021-03819-2. - DOI - PMC - PubMed

[2] Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A, et al., Highly accurate protein structure prediction with AlphaFold, Nature 596 (2021) 583–589. 10.1038/s41586-021-03819-2. - DOI - PMC - PubMed

[3] Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A, et al., Applying and improving AlphaFold at CASP14, Proteins 89 (2021) 1711–1721. 10.1002/prot.26257. - DOI - PMC - PubMed

[4] Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A, et al., Applying and improving AlphaFold at CASP14, Proteins 89 (2021) 1711–1721. 10.1002/prot.26257. - DOI - PMC - PubMed

[5] Senior AW, Evans R, Jumper J, Kirkpatrick J, Sifre L, Green T, Qin C, Zidek A, Nelson AWR, Bridgland A, et al., Improved protein structure prediction using potentials from deep learning, Nature 577 (2020) 706–710. 10.1038/s41586-019-1923-7. - DOI - PubMed

[6] Senior AW, Evans R, Jumper J, Kirkpatrick J, Sifre L, Green T, Qin C, Zidek A, Nelson AWR, Bridgland A, et al., Improved protein structure prediction using potentials from deep learning, Nature 577 (2020) 706–710. 10.1038/s41586-019-1923-7. - DOI - PubMed

[7] Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, Ronneberger O, Willmore L, Ballard AJ, Bambrick J, et al., Accurate structure prediction of biomolecular interactions with AlphaFold 3, Nature (2024) 10.1038/s41586-024-07487-w. - DOI - PMC - PubMed

[8] Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, Ronneberger O, Willmore L, Ballard AJ, Bambrick J, et al., Accurate structure prediction of biomolecular interactions with AlphaFold 3, Nature (2024) 10.1038/s41586-024-07487-w. - DOI - PMC - PubMed

[9] Ahdritz G, Bouatta N, Floristean C, Kadyan S, Xia Q, Gerecke W, O’Donnell TJ, Berenberg D, Fisk I, Zanichelli N, et al., OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization, Nat Methods (2024) 10.1038/s41592-024-02272-z. - DOI - PMC - PubMed

[10] Ahdritz G, Bouatta N, Floristean C, Kadyan S, Xia Q, Gerecke W, O’Donnell TJ, Berenberg D, Fisk I, Zanichelli N, et al., OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization, Nat Methods (2024) 10.1038/s41592-024-02272-z. - DOI - PMC - PubMed

Account

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction

Affiliations

Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction

Authors

Affiliations

Abstract

Conflict of interest statement

Update of

References

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources

Full Text Sources