Unmasking AlphaFold to integrate experiments and predictions in multimeric complexes
Open accessThe DOI matches its Crossref record; 17 of the 17 references it cites were found in a registry or an index.
Abstract
Since the release of AlphaFold, researchers have actively refined its predictions and attempted to integrate it into existing pipelines for determining protein structures. These efforts have introduced a number of functionalities and optimisations at the latest Critical Assessment of protein Structure Prediction edition (CASP15), resulting in a marked improvement in the prediction of multimeric protein structures. However, AlphaFold’s capability of predicting large protein complexes is still limited and integrating experimental data in the prediction pipeline is not straightforward. In this study, we introduce AF_unmasked to overcome these limitations. Our results demonstrate that AF_unmasked can integrate experimental information to build larger or hard to predict protein assemblies with high confidence. The resulting predictions can help interpret and augment experimental data. This approach generates high quality (DockQ score > 0.8) structures even when little to no evolutionary information is available and imperfect experimental structures are used as a starting point. AF_unmasked is developed and optimised to fill incomplete experimental structures (structural inpainting), which may provide insights into protein dynamics. In summary, AF_unmasked provides an easy-to-use method that efficiently integrates experiments to predict large protein complexes more confidently.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Highly accurate protein structure prediction with AlphaFold
Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03819-2 · Cited by 43,764
Accurate prediction of protein structures and interactions using a 3-track network
Baek, M., DiMaio, F., Anishchenko, I., Dauparas, J., Ovchinnikov, S., Lee, G. R., Wang, J., Cong, Q., Kinch, L. N., Schaeffer, R. D., et al. (2021) · doi:10.1101/2021.06.14.448402 · Cited by 5,549 · cited 1×
“The idea was later generalised to directly fi lling in gaps in 3D structures in RF joint 51 , a modi fi ed version of RoseTTAFold 52 weremaskingfragmentsofproteinsequence/structure pairs allows to predict native-like structures while generating novel sequences in the masked regions.”
BackgroundProtein complex prediction with AlphaFold-Multimer
Evans, R., O’Neill, M., Pritzel, A., Antropova, N., Senior, A., Green, T., Žídek, A., Bates, R., Blackwell, S., Yim, J., et al. (2021) · doi:10.1101/2021.10.04.463034 · Cited by 3,808 · cited 3×
“This type of complex is a good test case, as AlphaFold is relatively weak in modelling antigen-antibody interactions 3 , likely due to a lack of evolutionary coupling signal in the MSA 16 .”
BackgroundTM-align: a protein structure alignment algorithm based on the TM-score
Zhang, Y. (2005) · Nucleic Acids Research · doi:10.1093/nar/gki524 · Cited by 3,774
Fast and accurate protein structure search with Foldseek
van Kempen, M., Kim, S. S., Tumescheit, C., Mirdita, M., Lee, J., Gilchrist, C. L. M., Söding, J., & Steinegger, M. (2022) · doi:10.1101/2022.02.07.479398 · Cited by 2,598
Fast and accurate protein structure search with Foldseek
van Kempen, M., Kim, S. S., Tumescheit, C., Mirdita, M., Lee, J., Gilchrist, C. L. M., Söding, J., & Steinegger, M. (2024) · Nature Biotechnology · doi:10.1038/s41587-023-01773-0 · Cited by 2,596
Scaffolding protein functional sites using deep learning
Wang, J., Lisanza, S., Juergens, D., Tischer, D., Watson, J. L., Castro, K. M., Ragotte, R., Saragovi, A., Milles, L. F., Baek, M., et al. (2022) · Science · doi:10.1126/science.abn2100 · Cited by 487 · cited 1×
“The idea was later generalised to directly fi lling in gaps in 3D structures in RF joint 51 , a modi fi ed version of RoseTTAFold 52 weremaskingfragmentsofproteinsequence/structure pairs allows to predict native-like structures while generating novel sequences in the masked regions.”
BackgroundDockQ: A Quality Measure for Protein-Protein Docking Models
Basu, S., & Wallner, B. (2016) · PLOS ONE · doi:10.1371/journal.pone.0161879 · Cited by 460 · cited 1×
“To describe a structure prediction qualitatively,wealsofollowtheDockQqualityclassi fi cation,according to which a prediction can be: Incorrect (DockQ score below 0.23), Acceptable (DockQ between 0.23 and 0.49), Medium (DockQ between 0.49 and 0.8), High quality (DockQ score above 0.8) 19 .”
BackgroundInfluentialSelf-citationAhdritz, G., Bouatta, N., Floristean, C., Kadyan, S., Xia, Q., Gerecke, W., O’Donnell, T. J., Berenberg, D., Fisk, I., Zanichelli, N., et al. (2024) · Nature Methods · doi:10.1038/s41592-024-02272-z · Cited by 439 · cited 2×
“Other integrative approaches propose retraining versions of AlphaFold that are better at taking experiments into account: versions of OpenFold 12 and Uni-Fold 13 (retrainable implementations of Alpha-Fold)acceptextrainputdataderivedfromcrosslinkingexperimentsto guide monomeric and multimeric…”
MethodUsing AlphaFold to predict the impact of single mutations on protein stability and function
Pak, M. A., Markhieva, K. A., Novikova, M. S., Petrov, D. S., Vorobyev, I. S., Maksimova, E. S., Kondrashov, F. A., & Ivankov, D. N. (2023) · PLOS One · doi:10.1371/journal.pone.0282689 · Cited by 322 · cited 1×
“AlphaFold is not trained to predict the effect of mutations on the folding of a protein, and it cannot predict the impact of single-point mutations on protein stability 21 .”
Background
Research works printed but not linked: 41
- Bryant, P. et al. Predicting the structure of large protein complexes using Alphafold and Monte Carlo tree search. Nat. Commun. 13, 6028 (2022). · doi:10.1038/s41467-022-33729-4
- Terwilliger, T. C. et al. AlphaFold predictions are valuable hypotheses, and accelerate but do not replace experimental structure determination. bioRxiv https://doi.org/10.1101/2022.11.21.517405 (2022). · DOI · Paper page
- Liebschner, D. Using predicted models in Phenix. Acta Cryst. 75, 861–877 (2019).
- Terwilliger, T. C. et al. Improved AlphaFold modeling with implicit experimental information. Nat. Methods 19, 1376–1382 (2022). · DOI · Paper page
- Li, Z. et al. Uni-fold: An open-source platform for developing protein folding models beyond Alphafold. bioRxiv https://www.biorxiv.org/content/10.1101/2022.08.04.502811v3.full.pdf (2022). · DOI · Paper page
- Stahl, K., Brock, O. & Rappsilber, J. Modelling protein complexes with crosslinking mass spectrometry and deep learning. bioRxiv https://doi.org/10.1101/2023.06.07.544059 (2023). · DOI · Paper page
- Mirabello, C. & Wallner, B. DockQv2: Improved automatic quality measure forprotein multimers, nucleic acids and small molecules. Bioinformatics. btae586 https://doi.org/10.1093/bioinformatics/btae586 (2024). · DOI · Paper page
- Gray, J. J. et al. Protein–protein docking with simultaneous optimization of rigid-body displacement and side-chain conformations. J. Mol. Biol. 331, 281–299 (2003). · DOI · Paper page
- Mao, Y. et al. The small subunit of Rubisco and its potential as an engineering target. J. Exp. Bot. 74, 543–561 (2023). · DOI
- Valegård, K., Hasse, D., Andersson, I. & Gunn, L. H. Structure of rubisco from arabidopsis thaliana in complex with 2-carboxyarabinitol-1,5-bisphosphate. Acta Crystallogr. D. Struct. Biol. 74, 1–9 (2018). · DOI
- Yin, Y. et al. Structural basis for aggregate dissolution and refolding by the mycobacterium tuberculosis ClpB-DnaK bi-chaperone system. Cell Rep. 35, 109166 (2021). · DOI
- Kędzierska-Mieszkowska, S. & Zolkiewski, M. Hsp100 molecular chaperone clpb and its role in virulence of bacterial pathogens. Int. J. Mol. Sci. 22, 5319 (2021). · DOI
- Katikaridis, P., Bohl, V. & Mogk, A. Resisting the heat: bacterial disaggregases rescue cells from devastating protein aggregation. Front. Mol. Biosci. 8, 681439 (2021). · DOI
- Avellaneda, M. J. et al. Processive extrusion of polypeptide loops by a hsp100 disaggregase. Nature 578, 317–320 (2020). · DOI
- Mazal, H., Iljina, M., Riven, I. & Haran, G. Ultrafast pore-loop dynamics in a AAA+ machine point to a Brownian-ratchet mechanism for protein translocation. Sci. Adv. 7, eabg4674 (2021). · DOI
- Uchihashi, T. et al. Dynamic structural states of ClpB involved in its disaggregation function. Nat. Commun. 9, 2147 (2018). · doi:10.1038/s41467-018-04587-w
- Deville, C., Franke, K., Mogk, A., Bukau, B. & Saibil, H. R. Two-step activation mechanism of the ClpB disaggregase for sequential substrate threading by the main ATPase motor. Cell Rep. 27, 3433–3446.e4 (2019). · DOI
- Rizo, A. N. et al. Structural basis for substrate gripping and translocation by the ClpB AAA+ disaggregase. Nat. Commun. 10, 1–12 (2019). · DOI · Paper page
- Yu, H. et al. ATP hydrolysis-coupled peptide translocation mechanism of Mycobacterium tuberculosis ClpB. Proc. Natl Acad. Sci. USA 115, E9560–E9569 (2018). · DOI
- Deville, C. et al. Structural pathway of regulated substrate transfer and threading through an hsp100 disaggregase. Sci. Adv. 3, e1701726 (2017). · DOI · Paper page
- Haslberger, T. et al. M domains couple the ClpB threading motor with the DnaK chaperone activity. Mol. Cell 25, 247–260 (2007). · DOI · Paper page
- Oguchi, Y. et al. A tightly regulated molecular toggle controls AAA+ disaggregase. Nat. Struct. Mol. Biol. 19, 1338–1346 (2012). · DOI
- Rosenzweig, R., Moradi, S., Zarrine-Afsar, A., Glover, J. R. & Kay, L. E. Unraveling the mechanism of protein disaggregation through a ClpB-DnaK interaction. Science 339, 1080–1083 (2013). · DOI · Paper page
- Carroni, M. et al. Head-to-tail interactions of the coiled-coil domains regulate ClpB activity and cooperation with hsp70 in protein disaggregation. Elife 3, e02481 (2014). · DOI
- Mazal, H. et al. Tunable microsecond dynamics of an allosteric switch regulate the activity of a AAA+ disaggregation machine. Nat. Commun. 10, 1438 (2019). · DOI
- Rosenzweig, R. et al. Clpb n-terminal domain plays a regulatory role in protein disaggregation. Proc. Natl Acad. Sci. USA 112, E6872–E6881 (2015). · DOI
- Bergoug, M. et al. Neurofibromin structure, functions and regulation. Cells 9, 2365 (2020). · DOI · Paper page
- Gutmann, D. H., Wood, D. L. & Collins, F. S. Identification of the neurofibromatosis type 1 gene product. Proc. Natl Acad. Sci. USA 88, 9658–9662 (1991). · DOI · Paper page
- Ratner, N. & Miller, S. J. A RASopathy gene commonly mutated in cancer: the neurofibromatosis type 1 tumour suppressor. Nat. Rev. Cancer 15, 290–301 (2015). · DOI · Paper page
- Naschberger, A., Baradaran, R., Rupp, B. & Carroni, M. The structure of neurofibromin isoform 2 reveals different functional states. Nature 599, 315–319 (2021). · DOI
- Lupton, C. J. et al. The cryo-EM structure of the human neurofibromin dimer reveals the molecular basis for neurofibromatosis type 1. Nat. Struct. Mol. Biol. 28, 982–988 (2021). · DOI
- Chaker-Margot, M. et al. Structural basis of activation of the tumor suppressor protein neurofibromin. Mol. Cell 82, 1288–1296.e5 (2022). · DOI
- Young, L. C. et al. Destabilizing NF1 variants act in a dominant negative manner through neurofibromin dimerization. Proc. Natl Acad. Sci. USA 120, e2208960120 (2023). · DOI
- Li, Z., Nguyen, S. P., Xu, D. & Shang, Y. Protein loop modeling using deep generative adversarial network. In 2017 IEEE 29th International Conference on Tools with Artificial Intelligence (ICTAI). 1085–1091 (IEEE, 2017). · DOI · Paper page
- Anand, N. & Huang, P. Generative modeling for protein structures. Adv. Neural Inform. Process. Syst. https://doi.org/10.1038/s41587-023-02115-w (2018). · DOI
- Lai, B., McPartlon, M. & Xu, J. End-to-end deep structure generative model for protein design. bioRxiv https://doi.org/10.1101/2022.07.09.499440 (2022). · DOI
- Zhang, C. et al. Framedipt: Se (3) diffusion model for protein structure inpainting. bioRxiv https://doi.org/10.1101/2023.11.21.568057 (2023). · DOI
- Collins, K. W. et al. Dockground resource for protein recognition studies. Protein Sci. 31, e4481 (2022). · DOI
- Mukherjee, S. & Zhang, Y. Mm-align: a quick algorithm for aligning multiple-chain protein complex structures using iterative dynamic programming. Nucleic Acids Res. 37, e83–e83 (2009). · DOI · Paper page
- Zhang, Y. & Skolnick, J. Scoring function for automated assessment of protein structure template quality. Proteins Struct. Funct. Bioinform. 57, 702–710 (2004). · DOI · Paper page
- Pintilie, G. et al. Measurement of atom resolvability in cryo-em maps with q-scores. Nat. Methods 17, 328–334 (2020). · DOI · Paper page
Cited by
(2025) · doi:10.1016/j.cub.2025.06.020 · Cited by 25
AlphaPulldown2—a general pipeline for high-throughput structural modeling
Molodenskiy, D., Maurer, V. J., Yu, D., Chojnowski, G., Bienert, S., Tauriello, G., Gilep, K., Schwede, T., & Kosinski, J. (2025) · Bioinformatics · doi:10.1093/bioinformatics/btaf115 · Cited by 15
Estimating protein complex model accuracy using graph transformers and pairwise similarity graphs
Liu, J., Neupane, P., & Cheng, J. (2024) · Bioinformatics Advances · doi:10.1093/bioadv/vbaf180 · Cited by 6
Liu, L., Liu, Y., Ye, X., Zhang, S., Li, Y., Zhu, K., Xue, Y., Zhou, J., Zhang, X., & Fang, X. (2026) · Nature Machine Intelligence · doi:10.1038/s42256-026-01264-2 · Cited by 1
“Improving the accuracy of a single prediction has proved difficult, and generating large ensembles of candidate conformations has therefore emerged as a simple yet powerful way to boost performance.”
AlphaPulldown2 – A General Pipeline for High-Throughput Structural Modeling
Molodenskiy, D., Maurer, V. J., Yu, D., Chojnowski, G., Bienert, S., Tauriello, G., Gilep, K., Schwede, T., & Kosinski, J. (2024) · doi:10.1101/2024.11.28.625873
Metadata
- Authors
- Mirabello, C., Wallner, B., Nystedt, B., Azinas, S., & Carroni, M.
- Venue
- Nature Communications
- Published
- 2024-10-09
- Volume
- 15
- Issue
- 1
- Pages
- 8724-8724
- ISSN
- 2041-1723 (ISSN-L)
- OpenAlex
- W4403205338
- Cited by
- 46
- References
- 17
- FWCI
- 5.34
- Type
- journal article
- Language
- en
- Open access
- gold
- Licence
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Funders
- Science for Life Laboratory (BeyondFold)
- Knut och Alice Wallenbergs Stiftelse (KAW2021.0347)
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Unmasking AlphaFold to integrate experiments and predictions in multimeric complexes. https://app.cochat.ai/papers/10.1038/s41467-024-52951-w
- The paper
- Mirabello, C., Wallner, B., Nystedt, B., Azinas, S., & Carroni, M. (2024). Unmasking AlphaFold to integrate experiments and predictions in multimeric complexes. Nature Communications. https://doi.org/10.1038/s41467-024-52951-w
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.