OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization
Open accessPublished version of a preprint
Abstract
AlphaFold2 revolutionized structural biology with the ability to predict protein structures with exceptionally high accuracy. Its implementation, however, lacks the code and data required to train new models. These are necessary to (1) tackle new tasks, like protein-ligand complex structure prediction, (2) investigate the process by which the model learns and (3) assess the model's capacity to generalize to unseen regions of fold space. Here we report OpenFold, a fast, memory efficient and trainable implementation of AlphaFold2. We train OpenFold from scratch, matching the accuracy of AlphaFold2. Having established parity, we find that OpenFold is remarkably robust at generalizing even when the size and diversity of its training set is deliberately limited, including near-complete elisions of classes of secondary structure elements. By analyzing intermediate structures produced during training, we also gain insights into the hierarchical manner in which OpenFold learns to fold. In sum, our studies demonstrate the power and utility of OpenFold, which we believe will prove to be a crucial resource for the protein modeling community.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
ColabFold and OpenFold let users run a modified AF2 protocol on open-access servers within minutes.
“allow users to run a modified AF2 protocol on open access servers in minutes”
From The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins (2024) · cites this paper as [9]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · see citing works
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1× · see citing works
“We used Adam (Kingma and Ba 2015) with 𝛽 1 = 0.9 , 𝛽 2 = 0.99 , and 𝜖 = 10 −6 .”
InfluentialHighly 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 · cited 1× · see citing works
“11) has arguably solved the problem for natural proteins with sufficiently deep MSAs.”
PyTorch: An Imperative Style, High-Performance Deep Learning Library
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019) · arXiv · doi:10.48550/arxiv.1912.01703 · Cited by 31,473 · see citing works
Kabsch, W., & Sander, C. (1983) · Biopolymers · doi:10.1002/bip.360221211 · Cited by 15,320 · cited 9× · see citing works
“Regions corresponding to the beta sheets of Lsi1 are predicted by the mainly alpha model with approximately the right shape, except that their atomic coordinates are not sufficiently precise to enable DSSP to classify them as beta sheets.”
MethodInfluentialComparative Protein Modelling by Satisfaction of Spatial Restraints
Šali, A., & Blundell, T. L. (1993) · Journal of Molecular Biology · doi:10.1006/jmbi.1993.1626 · Cited by 12,977 · see citing works
ColabFold - Making protein folding accessible to all
Mirdita, M., Schütze, K., Moriwaki, Y., Heo, L., Ovchinnikov, S., & Steinegger, M. (2021) · doi:10.1101/2021.08.15.456425 · Cited by 9,391 · cited 1× · see citing works
“By analyzing intermediate structures produced during training, we also gain insights into the hierarchical manner in which OpenFold learns to fold.”
Physics-informed machine learning
Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021) · Nature Reviews Physics · doi:10.1038/s42254-021-00314-5 · Cited by 8,718 · cited 1× · see citing works
“For example, integrating physical priors into machine learning models is an area of outstanding scientific interest ([53]).”
Varadi, M., Anyango, S., Deshpande, M., Nair, S., Natassia, C., Yordanova, G., Yuan, D., Stroe, O., Wood, G., Laydon, A., et al. (2022) · Nucleic Acids Research · doi:10.1093/nar/gkab1061 · Cited by 7,850 · see citing works
Principles that Govern the Folding of Protein Chains
Anfinsen, C. B. (1973) · Science · doi:10.1126/science.181.4096.223 · Cited by 7,096 · cited 1× · see citing works
“Predicting protein structure from sequence has been a defining challenge of biology for decades 1,2 .”
I-TASSER: a unified platform for automated protein structure and function prediction
Roy, A., Küçükural, A., & Zhang, Y. (2010) · Nature Protocols · doi:10.1038/nprot.2010.5 · Cited by 6,093 · see citing works
Evolutionary-scale prediction of atomic level protein structure with a language model
Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., et al. (2022) · doi:10.1101/2022.07.20.500902 · Cited by 5,523 · cited 1× · see citing works
“These include the ESMFold method for protein structure prediction 50 , which replaces MSAs with protein language models 51–53 , and FastFold, a community effort that has implemented substantial improvements including fast model-parallel training and inference 41 .”
Protein 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 · see citing works
Improved protein structure prediction using potentials from deep learning
Senior, A. W., Evans, R., Jumper, J., Kirkpatrick, J., Sifre, L., Green, T., Qin, C., Žídek, A., Nelson, A. W. R., Bridgland, A., et al. (2020) · Nature · doi:10.1038/s41586-019-1923-7 · Cited by 3,378 · see citing works
Highly accurate protein structure prediction for the human proteome
Tunyasuvunakool, K., Adler, J., Wu, Z., Green, T., Zielinski, M., Žídek, A., Bridgland, A., Cowie, A., Meyer, C., Laydon, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03828-1 · Cited by 3,025 · cited 1× · see citing works
“It has also been used to predict the structures of hundreds of millions of proteins 19–21 .”
CATH – a hierarchic classification of protein domain structures
Orengo, C., Michie, A. D., Jones, S., Jones, D., Swindells, M. B., & Thornton, J. M. (1997) · Structure · doi:10.1016/s0969-2126(97)00260-8 · Cited by 2,608 · cited 6× · see citing works
“For this task, we use CATH, which assigns protein domains, in increasing order of specificity, to a class (C), architecture (A), topology (T) and homologous superfamily (H).”
InfluentialScoring function for automated assessment of protein structure template quality
Zhang, Y., & Skolnick, J. (2004) · Proteins: Structure, Function, and Bioinformatics · doi:10.1002/prot.20264 · Cited by 2,392 · cited 1× · see citing works
“AlphaFold was run with JAX version 0.3.13. https://doi.org/10.1038/s41592-024-02272-z”
InfluentialScaling Laws for Neural Language Models
Kaplan, J., McCandlish, S., Henighan, T., Brown, T. B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., & Amodei, D. (2020) · arXiv · Cited by 1,749 · see citing works
FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
Dao, T., Fu, D. Y., Ermon, S., Rudra, A., & Ré, C. (2022) · arXiv · doi:10.48550/arxiv.2205.14135 · Cited by 1,214 · cited 5× · see citing works
“As such, OpenFold can be readily used by the widest community of developers and interfaces with a rich ecosystem of existing machine learning software 26–29 .”
InfluentialHarnessing protein folding neural networks for peptide–protein docking
Tsaban, T., Varga, J. K., Avraham, O., Ben-Aharon, Z., Khramushin, A., & Schueler-Furman, O. (2022) · Nature Communications · doi:10.1038/s41467-021-27838-9 · Cited by 1,205 · see citing works
Dill, K. A., Ozkan, S. B., Shell, M. S., & Weikl, T. R. (2008) · Annual Review of Biophysics · doi:10.1146/annurev.biophys.37.092707.153558 · Cited by 1,163 · cited 1× · see citing works
“Predicting protein structure from sequence has been a defining challenge of biology for decades 1,2 .”
Protein Data Bank: the single global archive for 3D macromolecular structure data
wwPDB consortium, Burley, S. K., Berman, H. M., Bhikadiya, C., Bi, C., Chen, L., Costanzo, L. D., Christie, C., Duarte, J. M., Dutta, S., et al. (2019) · Nucleic Acids Research · doi:10.1093/nar/gky949 · Cited by 1,163 · see citing works
Protein 3D Structure Computed from Evolutionary Sequence Variation
Marks, D. S., Colwell, L. J., Sheridan, R., Hopf, T. A., Pagnani, A., Zecchina, R., & Sander, C. (2011) · PLoS ONE · doi:10.1371/journal.pone.0028766 · Cited by 1,144 · see citing works
Mariani, V., Biasini, M., Barbato, A., & Schwede, T. (2013) · Bioinformatics · doi:10.1093/bioinformatics/btt473 · Cited by 1,138 · cited 1× · see citing works
“During the initial fine-tuning and subsequent predicted TM fine-tuning phases, we manually sampled checkpoints at peaks in the validation lDDT-C α 33 .”
Unified rational protein engineering with sequence-based deep representation learning
Alley, E. C., Khimulya, G., Biswas, S., AlQuraishi, M., & Church, G. M. (2019) · Nature Methods · doi:10.1038/s41592-019-0598-1 · Cited by 1,131 · cited 1× · see citing works
“These include the ESMFold method for protein structure prediction 50 , which replaces MSAs with protein language models 51–53 , and FastFold, a community effort that has implemented substantial improvements including fast model-parallel training and inference 41 .”
Self-citationLGA: a method for finding 3D similarities in protein structures
Zemla, A. (2003) · Nucleic Acids Research · doi:10.1093/nar/gkg571 · Cited by 1,069 · cited 5× · see citing works
“This trend continues until the model is fully trained, where the gap between 10-residue fragments and whole proteins shrinks to a mere 10 GDT-TS points.”
MethodInfluentialImproved prediction of protein-protein interactions using AlphaFold2
Bryant, P., Pozzati, G., & Elofsson, A. (2021) · doi:10.1101/2021.09.15.460468 · Cited by 1,025 · cited 1× · see citing works
“Predicting protein structure from sequence has been a defining challenge of biology for decades 1,2 .”
Training Compute-Optimal Large Language Models
Hoffmann, J., Borgeaud, S., Mensch, A., Buchatskaya, E., Cai, T., Rutherford, E., Casas, D. D. L., Hendricks, L. A., Welbl, J., Clark, A., et al. (2022) · Advances in Neural Information Processing Systems 35 · doi:10.52202/068431-2176 · Cited by 932 · cited 1× · see citing works
“It is possible that, in earlier phases of training, OpenFold has already recovered aspects of secondary structure not recognized by DSSP on account of imprecise atom positioning.”
Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model
Wang, S., Sun, S., Li, Z., Zhang, R., & Xu, J. (2017) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1005324 · Cited by 931 · cited 1× · see citing works
“These are necessary to (1) tackle new tasks, like protein–ligand complex structure prediction, (2) investigate the process by which the model learns and (3) assess the model’s capacity to generalize to unseen regions of fold space.”
(2020) · doi:10.1109/sc41405.2020.00024 · Cited by 928 · see citing works
Self-citationRasley, J., Rajbhandari, S., Ruwase, O., & He, Y. (2020) · Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3394486.3406703 · Cited by 823 · see citing works
Self-citationUniclust databases of clustered and deeply annotated protein sequences and alignments
Mirdita, M., von den Driesch, L., Galiez, C., Martin, M. J., Söding, J., & Steinegger, M. (2016) · Nucleic Acids Research · doi:10.1093/nar/gkw1081 · Cited by 710 · see citing works
Predicting multiple conformations via sequence clustering and AlphaFold2
Wayment-Steele, H. K., Ojoawo, A., Otten, R., Apitz, J. M., Pitsawong, W., Hömberger, M., Ovchinnikov, S., Colwell, L., & Kern, D. (2024) · Nature · doi:10.1038/s41586-023-06832-9 · Cited by 545 · see citing works
CATH: increased structural coverage of functional space
Sillitoe, I., Bordin, N., Dawson, N., Waman, V. P., Ashford, P., Scholes, H. M., Pang, C. S. M., Woodridge, L., Rauer, C., Sen, N., et al. (2021) · Nucleic Acids Research · doi:10.1093/nar/gkaa1079 · Cited by 509 · see citing works
Single-sequence protein structure prediction using a language model and deep learning
Chowdhury, R., Bouatta, N., Biswas, S., Floristean, C., Kharkar, A., Roy, K., Rochereau, C., Ahdritz, G., Zhang, J., Church, G. M., et al. (2022) · Nature Biotechnology · doi:10.1038/s41587-022-01432-w · Cited by 455 · cited 1× · see citing works
“It is possible that, in earlier phases of training, OpenFold has already recovered aspects of secondary structure not recognized by DSSP on account of imprecise atom positioning.”
Self-citationSingle-sequence protein structure prediction using language models from deep learning
Chowdhury, R., Bouatta, N., Biswas, S., Rochereau, C., Church, G. M., Sorger, P. K., & AlQuraishi, M. (2021) · doi:10.1101/2021.08.02.454840 · Cited by 455 · see citing works
Self-citationHigh-resolution <i>de novo</i> structure prediction from primary sequence
Wu, R., Ding, F., Wang, R., Shen, R., Zhang, X., Luo, S., Su, C., Wu, Z., Xie, Q., Berger, B., et al. (2022) · doi:10.1101/2022.07.21.500999 · Cited by 392 · cited 1× · see citing works
“These include the ESMFold method for protein structure prediction ([54]), which replaces MSAs with protein language models ([55], [56], [57]), and FastFold, a community effort that has implemented significant improvements including fast model-parallel training and inference ([46]).”
Andreeva, A. E., Kulesha, E., Gough, J., & Murzin, A. G. (2019) · Nucleic Acids Research · doi:10.1093/nar/gkz1064 · Cited by 372 · cited 1× · see citing works
“Multiple structural taxonomies for proteins exist, including the hierarchical classification of protein domain structures (CATH) 34,35 and structural classification of proteins (SCOP) 36 classification systems.”
Accurate prediction of protein–nucleic acid complexes using RoseTTAFoldNA
Baek, M., McHugh, R., Anishchenko, I., Jiang, H., Baker, D., & DiMaio, F. (2024) · Nature Methods · doi:10.1038/s41592-023-02086-5 · Cited by 372 · see citing works
End-to-End Differentiable Learning of Protein Structure
AlQuraishi, M. (2019) · Cell Systems · doi:10.1016/j.cels.2019.03.006 · Cited by 371 · see citing works
Self-citationJones, D. T., Singh, T., Kościółek, T., & Tetchner, S. (2014) · Bioinformatics · doi:10.1093/bioinformatics/btu791 · Cited by 345 · cited 1× · see citing works
“Building on a line of work applying deep learning to coevolutionary information encoded in multiple-sequence alignments (MSAs) 3–8 and homologous structures 9,10 , AlphaFold2 (ref.”
State-of-the-Art Estimation of Protein Model Accuracy Using AlphaFold
Roney, J. P., & Ovchinnikov, S. (2022) · Physical Review Letters · doi:10.1103/physrevlett.129.238101 · Cited by 281 · see citing works
Falcon, W., Borovec, J., Eggert, N., Ir1dXD, Præsius, S., Bapat, S., Schröter, H., Singh, A., Bereznyuk, V., Fan, T. J., et al. (2019) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.3530844 · Cited by 256 · see citing works
Improved protein structure prediction by deep learning irrespective of co-evolution information
(2021) · doi:10.1038/s42256-021-00348-5 · Cited by 256 · cited 1× · see citing works
“Building on a line of work applying deep learning to coevolutionary information encoded in multiple-sequence alignments (MSAs) 3–8 and homologous structures 9,10 , AlphaFold2 (ref.”
Genomics-aided structure prediction
Sulkowska, J. I., Morcos, F., Weigt, M., Terence Hwa, B., & Onuchic, J. N. (2012) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1207864109 · Cited by 226 · cited 1× · see citing works
“MSAs implicitly encode a substantial amount of structural knowledge, as evidenced by early coevolution-based structure prediction methods that were entirely unsupervised, making no use of experimental structural data 45,46 .”
Protein structure similarities
Koehl, P. (2001) · Current Opinion in Structural Biology · doi:10.1016/s0959-440x(00)00214-1 · Cited by 176 · cited 1× · see citing works
“To investigate this learning hypothesis, we created 1D and 2D PCA projections of predicted structures at each timestep and compared them to the full 3D prediction at training step 5,000, at the end of the rapid rise in accuracy, using the translationally- and rotationallyinvariant distance-based root mean square devia…”
Haas, J., Barbato, A., Behringer, D., Studer, G., Roth, S., Bertoni, M., Mostaguir, K., Gumienny, R., & Schwede, T. (2018) · Proteins: Structure, Function, and Bioinformatics · doi:10.1002/prot.25431 · Cited by 168 · see citing works
Enhancing Evolutionary Couplings with Deep Convolutional Neural Networks
Liu, Y., Palmedo, P., Ye, Q., Berger, B., & Peng, J. (2018) · Cell Systems · doi:10.1016/j.cels.2017.11.014 · Cited by 114 · cited 1× · see citing works
“These are necessary to (1) tackle new tasks, like protein–ligand complex structure prediction, (2) investigate the process by which the model learns and (3) assess the model’s capacity to generalize to unseen regions of fold space.”
‘The entire protein universe’: AI predicts shape of nearly every known protein
Callaway, E. (2022) · Nature · doi:10.1038/d41586-022-02083-2 · Cited by 113 · cited 1× · see citing works
“The model has been made available to the public with DeepMind’s official open-source implementation, which has been used to predict the structures of hundreds of millions of proteins (Tunyasuvunakool et al. 2021, Varadi et al. 2021, Callaway 2022).”
MethodA backbone-centred energy function of neural networks for protein design
Huang, B., Xu, Y., Hu, X., Liu, Y., Liao, S., Zhang, J., Huang, C., Hong, J., Chen, Q., & Liu, H. (2022) · Nature · doi:10.1038/s41586-021-04383-5 · Cited by 103 · see citing works
Kernel Operations on the GPU, with Autodiff, without Memory OverflowsPreprint
Charlier, B., Feydy, J., Glaunès, J., Collin, F.-D., & Durif, G. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2004.11127 · Cited by 84 · cited 2× · see citing works
“Predictions made by OpenFold and AlphaFold2 on the CAMEO validation set are assessed using the lDDT-C α 33 (Local Distance Difference Test with respect to the alpha carbon) metric (Fig.”
Pearce, R., Omenn, G. S., & Zhang, Y. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.05.15.491755 · Cited by 71 · cited 1× · see citing works
“It is possible that, in earlier phases of training, OpenFold has already recovered aspects of secondary structure not recognized by DSSP on account of imprecise atom positioning.”
Structure of a B12-dependent radical SAM enzyme in carbapenem biosynthesis
Knox, H. L., Sinner, E. K., Townsend, C. A., Boal, A. K., & Booker, S. J. (2022) · Nature · doi:10.1038/s41586-021-04392-4 · Cited by 67 · see citing works
Prediction of multiple conformational states by combining sequence clustering with AlphaFold2Preprint
Wayment-Steele, H. K., Ovchinnikov, S., Colwell, L. J., & Kern, D. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.10.17.512570 · Cited by 66 · cited 1× · see citing works
“16–18).”
Structural basis for high selectivity of a rice silicon channel Lsi1
Saitoh, Y., Mitani‐Ueno, N., Saito, K., Matsuki, K., Huang, S., Yang, L., Yamaji, N., Ishikita, H., Shen, J., Feng, J. X., et al. (2021) · Nature Communications · doi:10.1038/s41467-021-26535-x · Cited by 61 · cited 1× · see citing works
“Because the topology ablations were evaluated on held-out topologies, the architecture ablations were evaluated on held-out architectures and so on, validation scores from different data eli-sion experiments cannot be compared directly.”
Accurate prediction of nucleic acid and protein-nucleic acid complexes using RoseTTAFoldNAPreprint
Baek, M., McHugh, R., Anishchenko, I. V., Baker, D. A., & DiMaio, F. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.09.09.507333 · Cited by 60 · cited 1× · see citing works
“…and to employ it as a module within novel algorithms, including ones for protein complex prediction (Baek 2021), peptide-protein interactions (Tsaban et al. 2022), structure ranking (Roney and Ovchinnikov 2022), and more ( e.g. , Baltzis et al. 2022, Bryant et al. 2022, Wayment-Steele et al. 2022).”
MethodUni-Fold: An Open-Source Platform for Developing Protein Folding Models beyond AlphaFoldPreprint
Li, Z., Liu, X., Chen, W., Shen, F., Bi, H., Ke, G., & Zhang, L. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.08.04.502811 · Cited by 59 · cited 2× · see citing works
“OpenFold, unlike AlphaFold2, was trained using per-batch as opposed to per-sample gradient clipping (first noted by the Uni-Fold team 42 ).”
An end-to-end deep learning method for protein side-chain packing and inverse folding
McPartlon, M., & Xu, J. (2023) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2216438120 · Cited by 55 · cited 1× · see citing works
“57,58).”
OneFlow: Redesign the Distributed Deep Learning Framework from ScratchPreprint
Yuan, J., Li, X., Cheng, C., Juncheng, L., Guo, R., Cai, S., Yao, C., Yang, F., Yi, X., Wu, C., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.15032 · Cited by 53 · see citing works
FastFold: Reducing AlphaFold Training Time from 11 Days to 67 HoursPreprint
Cheng, S., Zhao, X.-L., Lu, G.-Y., Fang, J.-R., Yu, Z.-M., Zheng, T.-Y., Wu, R., Zhang, X., Peng, J., & You, Y. (2022) · arXiv.org · doi:10.48550/arxiv.2203.00854 · Cited by 48 · see citing works
A comparison of reduced coordinate sets for describing protein structure
Hinsen, K., Hu, S., Kneller, G. R., & Niemi, A. J. (2013) · The Journal of Chemical Physics · doi:10.1063/1.4821598 · Cited by 39 · see citing works
Scaling Laws vs Model Architectures: How does Inductive Bias Influence Scaling?
Tay, Y., Dehghani, M., Abnar, S., Chung, H., Fedus, W., Rao, J., Narang, S., Tran, V., Yogatama, D., & Metzler, D. (2023) · Findings of the Association for Computational Linguistics: EMNLP 2023 · doi:10.18653/v1/2023.findings-emnlp.825 · Cited by 34 · see citing works
Highly significant improvement of protein sequence alignments with AlphaFold2
(2022) · doi:10.1093/bioinformatics/btac625 · Cited by 32 · cited 1× · see citing works
“16–18).”
Golkov, V., Skwark, M. J., Golkov, A., Dosovitskiy, A., Brox, T., Meiler, J., & Cremers, D. (2016) · neural information processing systems · Cited by 29 · see citing works
Caught in motion: human NTHL1 undergoes interdomain rearrangement necessary for catalysis
Carroll, B. L., Zahn, K. E., Hanley, J. P., Wallace, S. S., Dragon, J. A., & Doublié, S. (2021) · Nucleic Acids Research · doi:10.1093/nar/gkab1162 · Cited by 26 · see citing works
HelixFold: An Efficient Implementation of AlphaFold2 using PaddlePaddlePreprint
Wang, G., Fang, X., Wu, Z., Liu, Y., Xue, Y., Xiang, Y., Yu, D., Wang, F., & Ma, Y. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2207.05477 · Cited by 26 · see citing works
Singh, J., Paliwal, K. K., Litfin, T., Singh, J., & Zhou, Y. (2022) · Bioinformatics · doi:10.1093/bioinformatics/btac421 · Cited by 25 · cited 1× · see citing works
“54–56).”
Molecular basis of antibiotic self-resistance in a bee larvae pathogen
Dang, T., Loll, B., Müller, S., Skobalj, R., Ebeling, J., Bulatov, T. M., Gensel, S., Göbel, J., Wahl, M. C., Genersch, E., et al. (2022) · Nature Communications · doi:10.1038/s41467-022-29829-w · Cited by 22 · see citing works
Coexistence of Phases in a Protein Heterodimer
Krokhotin, A., Liwo, A., Niemi, A. J., & Scheraga, H. A. (2012) · The Journal of Chemical Physics · doi:10.1063/1.4734019 · Cited by 21 · see citing works
A Deep SE(3)-Equivariant Model for Learning Inverse Protein FoldingPreprint
McPartlon, M., Lai, B., & Xu, J. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.04.15.488492 · Cited by 18 · see citing works
Self-attention Does Not Need $O(n^2)$ Memory
Rabe, M. N., & Staats, C. (2021) · arXiv · doi:10.48550/arxiv.2112.05682 · Cited by 17 · cited 1× · see citing works
“…of neural network attention mechanisms (Vaswani et al. 2017) with favorable properties for unusually short and long sequences (Rabe and Staats 2021, Dao et al. 2022), module refactoring for lower memory usage, optional approximations of certain computations that reduce the memory…”
BackgroundOpenProteinSet: Training data for structural biology at scalePreprint
Ahdritz, G., Bouatta, N., Kadyan, S., Jarosch, L., Berenberg, D., Fisk, I., Watkins, A. M., Ra, S., Bonneau, R., & AlQuraishi, M. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2308.05326 · Cited by 15 · cited 2× · see citing works
“As the AlphaFold2 MSA database has not been publicly released, we used OpenProteinSet 23 , a replication of the AlphaFold2 training dataset that substitutes newer versions of sequence databases when available.”
Self-citationWei, X., Wang, Y., Wen, B., Liu, S., Wang, L., Sun, L., Gu, T.-Y., Li, Z., Bao, Y., Fan, S., et al. (2021) · Journal of Agricultural and Food Chemistry · doi:10.1021/acs.jafc.1c00940 · Cited by 14 · see citing works
Structural and thermodynamic analyses of human TMED1 (p24γ1) Golgi dynamics
Mota, D. C. A. M., Cardoso, I. A., Mori, R. M., Batista, M., Basso, L. G. M., Nonato, M. C., Costa-Filho, A. J., & Mendes, L. F. S. (2021) · Biochimie · doi:10.1016/j.biochi.2021.10.002 · Cited by 14 · see citing works
FastFold: Optimizing AlphaFold Training and Inference on GPU Clusters
Cheng, S., Zhao, X., Lu, G., Fang, J., Zheng, T., Wu, R., Zhang, X., Peng, J., & You, Y. (2024) · ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming · doi:10.1145/3627535.3638465 · Cited by 11 · cited 3× · see citing works
“OpenFold training speed matches or improves upon that of AlphaFold2, as has been reported by other researchers using OpenFold 41,42 , and it can be trained on long crop sizes (up to 1,200 in the ‘initial training’ setting).”
Scaling Laws vs Model Architectures: How does Inductive Bias Influence Scaling?
Tay, Y., Dehghani, M., Abnar, S., Chung, H. W., Fedus, W., Rao, J., Narang, S., Tran, V. Q., Yogatama, D., & Metzler, D. (2022) · arXiv · doi:10.48550/arxiv.2207.10551 · Cited by 11 · see citing works
An end-to-end deep learning method for rotamer-free protein side-chain packingPreprint
McPartlon, M., & Xu, J. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.03.11.483812 · Cited by 4 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Evolutionary-scale prediction of atomic level protein structure with a language model
Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., et al. (2022) · doi:10.1101/2022.07.20.500902 · Cited by 5,523 · see citing works
Varadi, M., Bertoni, D., Magana, P., Paramval, U., Pidruchna, I., Radhakrishnan, M., Tsenkov, M., Nair, S., Mirdita, M., Yeo, J., et al. (2024) · Nucleic Acids Research · doi:10.1093/nar/gkad1011 · Cited by 2,052 · see citing works
Boltz-1 Democratizing Biomolecular Interaction ModelingPreprint
Wohlwend, J., Corso, G., Passaro, S., Getz, N., Reveiz, M., Leidal, K., Swiderski, W., Atkinson, L., Portnoi, T., Chinn, I., et al. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.11.19.624167 · Cited by 535 · see citing works
“We follow the OpenFold chunking implementation for the Triangular Attention layer [Ahdritz et al., 2024], and extend it to the MSA Module, applying it at three critical points: the transition layers, pair-weighted average layers, and outer product layers.”
Scalable emulation of protein equilibrium ensembles with generative deep learning
Lewis, S., Hempel, T., Jiménez-Luna, J., Gastegger, M., Xie, Y., Foong, A. Y. K., Satorras, V. G., Abdin, O., Veeling, B. S., Zaporozhets, I., et al. (2025) · Science · doi:10.1126/science.adv9817 · Cited by 335 · see citing works
ChatGPT for shaping the future of dentistry: the potential of multi-modal large language model
Huang, H., Zheng, O., Wang, D., Yin, J., Wang, Z., Ding, S., Yin, H., Xu, C., Yang, R., Zheng, Q., et al. (2023) · International Journal of Oral Science · doi:10.1038/s41368-023-00239-y · Cited by 324 · see citing works
Modeling conformational states of proteins with AlphaFold
Sala, D., Engelberger, F., Mchaourab, H. S., & Meiler, J. (2023) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2023.102645 · Cited by 182 · see citing works
Data Generation for Machine Learning Interatomic Potentials and Beyond
Kulichenko, M., Nebgen, B. T., Lubbers, N., Smith, J. S., Barros, K. M., Allen, A. E. A., Habib, A., Shinkle, E., Fedik, N. S., Li, Y. W., et al. (2024) · Chemical Reviews · doi:10.1021/acs.chemrev.4c00572 · Cited by 187 · see citing works
Protein generation with evolutionary diffusion: sequence is all you need
Alamdari, S., Thakkar, N., van den Berg, R., Tenenholtz, N., Strome, R., Moses, A. M., Lu, A. X., Fusi, N., Amini, A. P., & Yang, K. K. (2023) · doi:10.1101/2023.09.11.556673 · Cited by 154 · see citing works
“Multiple sequence alignment (MSA) EvoDiff models were trained on OpenFold [31], which contains 401,381 MSAs for 140,000 unique Protein Data Bank (PDB) chains and 16,000,000 UniClust30 clusters.”
MethodEasy and accurate protein structure prediction using ColabFold
Kim, G., Lee, S., Karin, E. L., Kim, H., Moriwaki, Y., Ovchinnikov, S., Steinegger, M., & Mirdita, M. (2024) · Nature Protocols · doi:10.1038/s41596-024-01060-5 · Cited by 167 · see citing works
Protein structure prediction with in-cell photo-crosslinking mass spectrometry and deep learning
Stahl, K., Graziadei, A., Dau, T., Brock, O., & Rappsilber, J. (2023) · Nature Biotechnology · doi:10.1038/s41587-023-01704-z · Cited by 145 · see citing works
“We use OpenFold v0.”
InfluentialAlphaFold predictions of fold-switched conformations are driven by structure memorization
Chakravarty, D., Schafer, J. W., Chen, E. A., Thole, J. F., Ronish, L. A., Lee, M., & Porter, L. L. (2024) · Nature Communications · doi:10.1038/s41467-024-51801-z · Cited by 159 · see citing works
“…contacts medium gray; coevolutionary information inferred from MSA using ACE, (teal). c Both AF2 multimer and AF2 predict the correctXCL1 dimerstructurefrom single sequences and 0 recycles, suggesting that they memorized its structure during training. match AF2 ’ s predictive performance 27 .”
InfluentialState-specific protein–ligand complex structure prediction with a multiscale deep generative model
Qiao, Z., Nie, W., Vahdat, A., Miller, T. F., & Anandkumar, A. (2024) · Nature Machine Intelligence · doi:10.1038/s42256-024-00792-z · Cited by 178 · see citing works
Efficient and accurate prediction of protein structure using RoseTTAFold2Preprint
Baek, M., Anishchenko, I. V., Humphreys, I. R., Cong, Q., Baker, D. A., & DiMaio, F. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.05.24.542179 · Cited by 152 · see citing works
The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins
Agarwal, V., & McShan, A. C. (2024) · Nature Chemical Biology · doi:10.1038/s41589-024-01638-w · Cited by 155 · see citing works
Yin, R., & Pierce, B. G. (2024) · Protein Science · doi:10.1002/pro.4865 · Cited by 145 · see citing works
AI-Driven Deep Learning Techniques in Protein Structure Prediction
Chen, L., Li, Q., Nasif, K. F. A., Xie, Y., Deng, B., Niu, S., Pouriyeh, S., Dai, Z., Chen, J., & Xie, Y. (2024) · International Journal of Molecular Sciences · doi:10.3390/ijms25158426 · Cited by 147 · see citing works
“OpenComplex is based on AlphaFold2 [25] and OpenFold [108].”
MethodProtenix - Advancing Structure Prediction Through a Comprehensive AlphaFold3 ReproductionPreprint
ByteDance AML AI4Science Team, Chen, X., Zhang, Y., Lu, C., Ma, W., Guan, J., Gong, C., Yang, J., Zhang, H., Zhang, K., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.01.08.631967 · Cited by 150 · see citing works
Structure prediction of alternative protein conformations
Bryant, P., & Noé, F. (2024) · Nature Communications · doi:10.1038/s41467-024-51507-2 · Cited by 105 · see citing works
AlphaFold prediction of structural ensembles of disordered proteins
Brotzakis, Z. F., Zhang, S., Murtada, M. H., & Vendruscolo, M. (2025) · Nature Communications · doi:10.1038/s41467-025-56572-9 · Cited by 99 · see citing works
Self-citationAFsample2 predicts multiple conformations and ensembles with AlphaFold2
Kalakoti, Y., & Wallner, B. (2025) · Communications Biology · doi:10.1038/s42003-025-07791-9 · Cited by 96 · see citing works
“Theideaisnotlimitedto theAF2inferencesystembutcouldbereadilyappliedtootherMSA-based predictionsystems,suchasRosettaFold 9 , OpenFold 25 , or AF3 26 .”
Progress at protein structure prediction, as seen in CASP15
Elofsson, A. (2023) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2023.102594 · Cited by 89 · see citing works
Qiu, X., Li, H., Steeg, G. V., & Godzik, A. (2024) · Biomolecules · doi:10.3390/biom14030339 · Cited by 100 · see citing works
“The AF2 breakthrough has been quickly followed by other AI tools such as RoseTTAfold [7], ESMFold [8], and OpenFold [9].”
BackgroundMethodKoh, H. Y., Zheng, Y., Yang, M., Arora, R., Webb, G. I., Pan, S., Li, L., & Church, G. M. (2025) · Nature Reviews Bioengineering · doi:10.1038/s44222-025-00349-8 · Cited by 96 · see citing works
Bartolec, T. K., Vázquez-Campos, X., Norman, A. R., Luong, C., Johnson, M., Payne, R. J., Wilkins, M. R., Mackay, J. P., & Low, J. K. K. (2023) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2219418120 · Cited by 81 · see citing works
“Therefore, development of more computationally efficient structural prediction software [e.g., OpenFold (79)], will be required to enable the prediction of missing subsets of the structural proteome and interactome.”
Deep language models for interpretative and predictive materials science
Hu, Y., & Buehler, M. J. (2023) · APL Machine Learning · doi:10.1063/5.0134317 · Cited by 80 · see citing works
Cai, H., Shen, C., Jian, T., Zhang, X., Chen, T., Han, X., Yang, Z., Dang, W., Hsieh, C., Kang, Y., et al. (2023) · Chemical Science · doi:10.1039/d3sc05552c · Cited by 77 · see citing works
Proteins with alternative folds reveal blind spots in AlphaFold-based protein structure prediction
Chakravarty, D., Lee, M., & Porter, L. L. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2024.102973 · Cited by 74 · see citing works
Geometric deep learning of protein–DNA binding specificity
Mitra, R., Li, J., Sagendorf, J. M., Jiang, Y., Cohen, A. S., Chiu, T.-P., Glasscock, C. J., & Rohs, R. (2024) · Nature Methods · doi:10.1038/s41592-024-02372-w · Cited by 75 · see citing works
“The rapid advancement of protein structure prediction meth-ods, including AlphaFold 26 , OpenFold 27 and RoseTTAFold 28 , along with protein–DNA complex modelers, such as RoseTTAFoldNA (RFNA) 29 , RoseTTAFold All-Atom 30 , MELD-DNA 31 and AlphaFold3 (ref.”
MethodNovel machine learning approaches revolutionize protein knowledge
Bordin, N., Dallago, C., Heinzinger, M., Kim, S., Littmann, M., Rauer, C., Steinegger, M., Rost, B., & Orengo, C. (2023) · Trends in Biochemical Sciences · doi:10.1016/j.tibs.2022.11.001 · Cited by 72 · see citing works
“DeepMind ’ s AF2 outperformed others in CASP14 [26], and reports suggest that high-quality models can be comparable to crystallographic structures, with competing methods reproducing DeepMind ’ s results [24,27,28].”
ResultDeep learning for protein structure prediction and design—progress and applications
Jänes, J., & Beltrão, P. (2024) · Molecular Systems Biology · doi:10.1038/s44320-024-00016-x · Cited by 71 · see citing works
“Finally, OpenFold (Ahdritz et al, 2022) is a reimplementation of AlphaFold2 that has the same architecture but having the code available as well as the data required for re-training.”
BackgroundMethodInfluentialTheoretical and Data-Driven Approaches for Biomolecular Condensates
Saar, K. L., Qian, D., Good, L. L., Morgunov, A. S., Collepardo‐Guevara, R., Best, R. B., & Knowles, T. P. J. (2023) · Chemical Reviews · doi:10.1021/acs.chemrev.2c00586 · Cited by 66 · see citing works
Casadevall, G., Duran, C., & Osuna, S. (2023) · JACS Au · doi:10.1021/jacsau.3c00188 · Cited by 68 · see citing works
FoldScript: a web server for the efficient analysis of AI-generated 3D protein models
Robert, X., Guillon, C., & Gouet, P. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkaf326 · Cited by 64 · see citing works
“This server allows to quickly and automatically analyse a set of models generated by AF3 and other AI-driven predictors (AF2-multimer, RoseTTAFold, ESMfold [ 7 ], OpenFold [ 8 ], etc.).”
Adaptive Sampling Methods for Molecular Dynamics in the Era of Machine Learning
Kleiman, D. E., Nadeem, H., & Shukla, D. (2023) · The Journal of Physical Chemistry B · doi:10.1021/acs.jpcb.3c04843 · Cited by 62 · see citing works
Transferable deep generative modeling of intrinsically disordered protein conformations
Janson, G., & Feig, M. (2024) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1012144 · Cited by 62 · see citing works
“This is in line with what has been observed for the OpenFold model [46]: protein modeling accuracy increases rapidly with the number of training systems and then continues to slowly improve by providing more systems.”
ResultDiaz, D. J., Gong, C., Ouyang-Zhang, J., Loy, J. M., Wells, J., Yang, D. K., Ellington, A. D., Dimakis, A. G., & Klivans, A. R. (2024) · Nature Communications · doi:10.1038/s41467-024-49780-2 · Cited by 66 · see citing works
“Deep learning is currently revolutionizing many physical and biological disciplines 17 – 21 , with AlphaFoldV2 leading the way as the scienti fi c breakthrough of the year in 2021 22,23 and sparking an entire waveofdeeplearningstructurepredictiontools 24,25 .”
Artificial intelligence in peptide-based drug design
Zhai, S., Liu, T., Lin, S., Li, D., Liu, H., Yao, X., & Hou, T. (2025) · Drug Discovery Today · doi:10.1016/j.drudis.2025.104300 · Cited by 66 · see citing works
actifpTM: a refined confidence metric of AlphaFold2 predictions involving flexible regions
Varga, J. K., Ovchinnikov, S., & Schueler‐Furman, O. (2025) · Bioinformatics · doi:10.1093/bioinformatics/btaf107 · Cited by 56 · see citing works
“Since 2020, AlphaFold2 [AF2 (Evans et al. 2021, Jumper et al. 2021)] and other similar deep learning algorithms (Baek et al. 2023, Ahdritz et al. 2024) have revolutionized structural biology, by making swift and confident predictions of millions of protein structures and complexes available.”
Merizo: a rapid and accurate protein domain segmentation method using invariant point attention
Lau, A. M., Kandathil, S. M., & Jones, D. T. (2023) · Nature Communications · doi:10.1038/s41467-023-43934-4 · Cited by 48 · see citing works
Overview of AlphaFold2 and breakthroughs in overcoming its limitations
Wang, L., Wen, Z., Liu, S.-W., Zhang, L., Finley, C., Lee, H., & Fan, H. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.108620 · Cited by 53 · see citing works
Large language models for drug discovery and development
Zheng, Y., Koh, H. Y., Ju, J., Yang, M., May, L. T., Webb, G. I., Li, L., Pan, S., & Church, G. M. (2025) · Patterns · doi:10.1016/j.patter.2025.101346 · Cited by 89 · see citing works
Strengthening nucleic acid biosecurity screening against generative protein design tools
Wittmann, B. J., Alexanian, T., Bartling, C. M., Beal, J., Clore, A. J., Diggans, J., Flyangolts, K., Gemler, B. T., Mitchell, T., Murphy, S. T., et al. (2025) · Science · doi:10.1126/science.adu8578 · Cited by 72 · see citing works
Are protein language models the new universal key?
Weißenow, K., & Rost, B. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.102997 · Cited by 51 · see citing works
Unmasking AlphaFold to integrate experiments and predictions in multimeric complexes
Mirabello, C., Wallner, B., Nystedt, B., Azinas, S., & Carroni, M. (2024) · Nature Communications · doi:10.1038/s41467-024-52951-w · Cited by 50 · see citing works
“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…”
MethodAlphaBridge: Tools for the analysis of predicted biomolecular complexes
Salmoral, D. Á., Borza, R., Maiella, C., Kwee, B. P. Y., Xie, R., Joosten, R. P., Hekkelman, M. L., & Anastassis Perrakis, A. (2026) · Structure · doi:10.1016/j.str.2026.08.011 · Cited by 48 · see citing works
An end-to-end framework for the prediction of protein structure and fitness from single sequence
Chen, Y., Xu, Y., Liu, D., Xing, Y., & Gong, H. (2024) · Nature Communications · doi:10.1038/s41467-024-51776-x · Cited by 45 · see citing works
Protein–protein interaction prediction methods: from docking-based to AI-based approaches
Tsuchiya, Y., Yamamori, Y., & Tomii, K. (2022) · Biophysical Reviews · doi:10.1007/s12551-022-01032-7 · Cited by 44 · see citing works
“Among the follow-up replicative studies of AF2, such as OpenFold (Ahdritz et al. 2022), Uni-Fold (Li et al. 2022a, b), MegaFold (Liu et al. 2022), and HelixFold (Wang et al. 2022), it was only Uni-Fold and Uni-Fold-symmetry (Li et al. 2022b) that could succeed not only in monomer prediction but also in multimer predic…”
MethodInfluentialEnd-to-end protein–ligand complex structure generation with diffusion-based generative models
Nakata, S., Mori, Y., & Tanaka, S. (2023) · BMC Bioinformatics · doi:10.1186/s12859-023-05354-5 · Cited by 43 · see citing works
“These methods are regression-based and rely on ad hoc techniques like MSA subsampling, changing the number of recycles, and enabling dropout during inference to sample multiple structures [44].”
BackgroundMethodLiu, J., Wu, L., Xie, A., Liu, W. L., Zhao, H., Wan, Y., & Mao, W. (2025) · Journal of Nanobiotechnology · doi:10.1186/s12951-025-03169-5 · Cited by 41 · see citing works
“Protein and antibody structure prediction has been revolutionized by deep learning-based tools such as AlphaFold2 [178], RoseTTAFold [209], OmegaFold [210], ESMFold [211], OpenFold [212], IgFold [213] and ABodyBuilder3 [214].”
Jin, S., Zeng, Z., Xiong, X., Huang, B., Tang, L., Wang, H., Ma, X., Tang, X., Shao, G., Huang, X., et al. (2025) · Communications Biology · doi:10.1038/s42003-025-08282-7 · Cited by 41 · see citing works
Brown, B. P., Stein, R. A., Meiler, J., & Mchaourab, H. S. (2024) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.3c01081 · Cited by 41 · see citing works
ABodyBuilder3: improved and scalable antibody structure predictions
Kenlay, H., Dreyer, F. A., Cutting, D., Nissley, D. A., & Deane, C. M. (2024) · Bioinformatics · doi:10.1093/bioinformatics/btae576 · Cited by 39 · see citing works
“We develop an efficient and scalable implementation of the ABodyBuilder2 architecture which makes use of vectorization to improve hardware utilization, leveraging optimisations from the OpenFold project (Ahdritz et al. 2024).”
Challenges in bridging the gap between protein structure prediction and functional interpretation
Varadi, M., Tsenkov, M., & Velankar, S. (2025) · Proteins: Structure, Function, and Bioinformatics · doi:10.1002/prot.26614 · Cited by 36 · see citing works
Aranganathan, A., Gu, X.-Y., Wang, D.-D., Vani, B. P., & Tiwary, P. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.103000 · Cited by 37 · see citing works
“From a data-based methods perspective, with the availability of an enormous structural database, particularly for proteins, methods such as AlphaFold2 (AF2) [7], RoseTTAFold2 [8], and OpenFold [9], have been developed to solve the protein folding problem.”
On the design space between molecular mechanics and machine learning force fields
Wang, Y., Takaba, K., Chen, M. S., Wieder, M., Xu, Y., Zhu, T., Zhang, J. Z. H., Nagle, A., Yu, K., Wang, X., et al. (2025) · Applied Physics Reviews · doi:10.1063/5.0237876 · Cited by 39 · see citing works
“It has quickly emerged to become a focus of cutting-edge research to generate samples directly on the coordinate space for biomolecular systems of interest— from small molecule conformer generation 272–276 , folded protein structure prediction 277–279 , to trajectory forecasting 30,280 .”
Representing structures of the multiple conformational states of proteins
Ramelot, T. A., Tejero, R., & Montelione, G. T. (2023) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2023.102703 · Cited by 26 · see citing works
“In particular, as the machine learning (ML) methods of AlphaFold2 (AF2) [2], RosettaFold [3], OpenFold [4], ESMFold [5], RaptorX [6], and other advanced techniques have reached the stage where single structure prediction of small proteins is robust and reliable, and a current frontier is…”
MethodProtein representations: Encoding biological information for machine learning in biocatalysis
Harding-Larsen, D., Funk, J., Madsen, N. G., Gharabli, H., Acevedo-Rocha, C. G., Mazurenko, S., & Welner, D. H. (2024) · Biotechnology Advances · doi:10.1016/j.biotechadv.2024.108459 · Cited by 38 · see citing works
AlphaFold and Protein Folding: Not Dead Yet! The Frontier Is Conformational Ensembles
Bowman, G. R. (2024) · Annual Review of Biomedical Data Science · doi:10.1146/annurev-biodatasci-102423-011435 · Cited by 35 · see citing works
Akpinaroglu, D., Seki, K., Guo, A., Zhu, E., Kelly, M. J. S., & Kortemme, T. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.12.15.571823 · Cited by 32 · see citing works
Natural Language Processing Methods for the Study of Protein–Ligand Interactions
Michels, J., Bandarupalli, R., Akbari, A. A., Le, T., Xiao, H. Y., Li, J., & Hom, E. F. Y. (2025) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.4c01907 · Cited by 33 · see citing works
AI-guided pipeline for protein–protein interaction drug discovery identifies a SARS-CoV-2 inhibitor
Trepte, P., Secker, C., Olivet, J., Blavier, J., Kostova, S., Maseko, S. B., Minia, I., Silva Ramos, E., Cassonnet, P., Golusik, S., et al. (2024) · Molecular Systems Biology · doi:10.1038/s44320-024-00019-8 · Cited by 35 · see citing works
“While we were limited to structure prediction-based PPI mapping for complexes with up to 1400 amino acids, further improvements in the structure prediction algorithms regarding speed, memory usage and the maximal size of proteins and protein complexes are already available (Ahdritz et al, 2022).”
Deep learning methods for protein structure prediction
Qin, Y., Chen, Z., Ye, P., Xiao, Y., Zhong, T., & Yu, X. (2024) · MedComm – Future Medicine · doi:10.1002/mef2.96 · Cited by 29 · see citing works
Mukherjee, S., Moafinejad, S. N., Badepally, N. G., Merdas, K., & Bujnicki, J. M. (2024) · Structure · doi:10.1016/j.str.2024.08.015 · Cited by 29 · see citing works
NMR investigations of glycan conformation, dynamics, and interactions
Angulo, J., Ardá, A., Bertuzzi, S., Canales, Á., Ereño‐Orbea, J., Gimeno, A., Gómez‐Redondo, M., Muñoz-García, J. C., Oquist-Phillips, P., Monaco, S., et al. (2024) · Progress in Nuclear Magnetic Resonance Spectroscopy · doi:10.1016/j.pnmrs.2024.10.002 · Cited by 29 · see citing works
Leveraging protein structural information to improve variant effect prediction
Gerasimavicius, L., Teichmann, S. A., & Marsh, J. A. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.103023 · Cited by 32 · see citing works
Wu, T., Stein, R. A., Kao, T., Brown, B. P., & Mchaourab, H. S. (2025) · Nature Communications · doi:10.1038/s41467-025-62582-4 · Cited by 33 · see citing works
“To fully leverage the potential of DEER data in protein structure prediction, we have developed DEERFold, a framework that directly incorporates spin label distance distributions derived from DEER data into AlphaFold2 through OpenFold 52 .”
InfluentialEllaway, J. I. J., Anyango, S., Nair, S. S., Zaki, H. A., Nadzirin, N., Powell, H. R., Gutmanas, A., Váradi, M., & Velankar, S. S. (2024) · Structural Dynamics · doi:10.1063/4.0000251 · Cited by 27 · see citing works
Artificial intelligence methods for protein folding and design
Zhang, Z., Ou, C., Cho, Y., Akiyama, Y., & Ovchinnikov, S. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.103066 · Cited by 27 · see citing works
Self-citationArtificial intelligence-driven computational methods for antibody design and optimization
Vecchietti, L. F., Wijaya, B. N., Armanuly, A., Hangeldiyev, B., Jung, H., Lee, S., Cha, M., & Kim, H. M. (2025) · mAbs · doi:10.1080/19420862.2025.2528902 · Cited by 28 · see citing works
“158 Moreover, methods like TAP 159 and PropertyDAG 160 have been developed to predict CDR-related developability properties and rank design candidates, respectively.”
Metamorphic proteins and how to find them
Porter, L. L., Artsimovitch, I., & Ramírez‐Sarmiento, C. A. (2024) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2024.102807 · Cited by 26 · see citing works
Bedraoui, A., Suntravat, M., Mejjad, S. E., Enezari, S., Oukkache, N., Sanchez, E. E., Galan, J. A., Fatimy, R. E., & Daouda, T. (2023) · Medicine in Drug Discovery · doi:10.1016/j.medidd.2023.100175 · Cited by 28 · see citing works
Heinzinger, M., & Rost, B. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.102986 · Cited by 25 · see citing works
Aubel, M., Eicholt, L. A., & Bornberg‐Bauer, E. (2023) · F1000Research · doi:10.12688/f1000research.130443.1 · Cited by 24 · see citing works
Structure-based prediction of protein-nucleic acid binding using graph neural networks
Sagendorf, J. M., Mitra, R., Huang, J., Chen, X. S., & Rohs, R. (2024) · Biophysical Reviews · doi:10.1007/s12551-024-01201-w · Cited by 25 · see citing works
“In addition to experimentally determined structures, advances in protein structure prediction (Baek et al. 2021; Jumper et al. 2021; Ahdritz et al. 2024) offer vast amounts of accurately predicted protein structures that can be analyzed with computational methods.”
Reading the repertoire: Progress in adaptive immune receptor analysis using machine learning
O’Donnell, T. J., Kanduri, C., Isacchini, G., Limenitakis, J. P., Brachman, R. A., Alvarez, R. A., Haff, I. H., Sandve, G. K., & Greiff, V. (2024) · Cell Systems · doi:10.1016/j.cels.2024.11.006 · Cited by 22 · see citing works
Self-citationAdvanced Strategies in Phage Research: Innovations, Applications, and Challenges
Wu, P., Li, W., Zhang, W., Li, S., Deng, B., Xu, S., & Li, Z. (2025) · Microorganisms · doi:10.3390/microorganisms13081960 · Cited by 25 · see citing works
“Structure prediction platforms such as AlphaFold and OpenFold have resolved previously uncharacterized proteins, including endolysins and tail fibers, enabling residue-level engineering [16,17].”
InfluentialDeep generative modeling of temperature-dependent structural ensembles of proteins
Janson, G., Jussupow, A., & Feig, M. (2025) · Communications Chemistry · doi:10.1038/s42004-025-01737-2 · Cited by 24 · see citing works
“The decoderisbasedontheStructureModulefromAF2,with fi veblocks 1,64 .”
Target identification and assessment in the era of AI
Pun, F. W., Podolskiy, D. I., Izumchenko, E., Mortlock, A. A., Oprea, T. I., Scheibye‐Knudsen, M., Fortney, K., Morgen, E. K., Ren, F., & Zhavoronkov, A. A. (2026) · Nature Reviews Drug Discovery · doi:10.1038/s41573-026-01412-8 · Cited by 28 · see citing works
“Partly in response to AlphaFold2, the OpenFold Consortium developed OpenFold to provide open access to its training code and data, while achieving comparable prediction accuracy 158 .”
Random, de novo, and conserved proteins: How structure and disorder predictors perform differently
Middendorf, L., & Eicholt, L. A. (2024) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.26652 · Cited by 24 · see citing works
“34,35 35 – 37 Being based on the structure module of AF2, all modern structure predictors provide a per residue confidence score established with AF2; pLDDT (predicted local distance difference test).”
BackgroundIntegrating AlphaFold and deep learning for atomistic interpretation of cryo-EM maps
Dai, X., Wu, L., Yoo, S., & Liu, Q. (2023) · Briefings in Bioinformatics · doi:10.1093/bib/bbad405 · Cited by 22 · see citing works
MHC-Fine: Fine-tuned AlphaFold for precise MHC-peptide complex prediction
Glukhov, E., Kalitin, D., Stepanenko, D., Zhu, Y., Nguyen, T., Jones, G. S., Patsahan, T., Simmerling, C. L., Mitchell, J. C., Vajda, S., et al. (2024) · Biophysical Journal · doi:10.1016/j.bpj.2024.05.011 · Cited by 22 · see citing works
Raisinghani, N., Alshahrani, M., Gupta, G., Tian, H., Xiao, S., Tao, P., & Verkhivker, G. M. (2024) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.4c00222 · Cited by 17 · see citing works
AI meets physics in computational structure-based drug discovery for GPCRs
Michino, M., Vendôme, J., & Kufareva, I. (2025) · npj Drug Discovery. · doi:10.1038/s44386-025-00019-0 · Cited by 25 · see citing works
Mieres‐Pérez, J., Almeida‐Hernández, Y., Sander, W., & Sánchez‐García, E. (2025) · Chemical Reviews · doi:10.1021/acs.chemrev.4c00807 · Cited by 21 · see citing works
Song, C., Gu, J., Ren, H., Li, Z., Xie, D., Han, L., Qiao, J., Cheng, Z., Nie, Y., & Zhou, Z. (2025) · Biotechnology Advances · doi:10.1016/j.biotechadv.2025.108716 · Cited by 21 · see citing works
Agentic AI and the rise of in silico team science in biomedical research
Li, B., Saini, A. K., Hernandez, J. G., & Moore, J. H. (2026) · Nature Biotechnology · doi:10.1038/s41587-026-03035-1 · Cited by 23 · see citing works
Large Language Models in Genomics—A Perspective on Personalized Medicine
Ali, S. M., Qadri, Y. A., Ahmad, K., Lin, Z., Leung, M.-F., Kim, S. W., Vasilakos, A. V., & Zhou, T. (2025) · Bioengineering · doi:10.3390/bioengineering12050440 · Cited by 23 · see citing works
Deep Learning for Flexible and Site-Specific Protein Docking and DesignPreprint
McPartlon, M., & Xu, J. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.04.01.535079 · Cited by 21 · see citing works
ESMFold Hallucinates Native-Like Protein SequencesPreprint
Jeliazkov, J. R., del Alamo, D., & Karpiak, J. D. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.05.23.541774 · Cited by 21 · see citing works
“Our approach to enabling design of input sequences in the ESMFold network, which is built on OpenFold [15] and is outlined in Figure 1, is no different: design is achieved by minimizing the loss gradient with respect to the input rather than to the weights.”
MethodModelling protein complexes with crosslinking mass spectrometry and deep learning
Stahl, K., Demann, L., Bremenkamp, R., Warneke, R., Hormes, B., Stülke, J., Brock, O., & Rappsilber, J. (2023) · doi:10.1101/2023.06.07.544059 · Cited by 20 · see citing works
The outcast of medicine: metals in medicine--from traditional mineral medicine to metallodrugs
Bai, D., Nowak, M., Lu, D., Wang, Q., Fitzgerald, M., Zhang, H., MacDonald, R., Xu, Z., & Luo, L. (2025) · Frontiers in Pharmacology · doi:10.3389/fphar.2025.1542560 · Cited by 20 · see citing works
Zhang, Y., Zhang, Z., Kagaya, Y., Terashi, G., Zhao, B., Xiong, Y., & Kihara, D. (2025) · Communications Biology · doi:10.1038/s42003-025-08783-5 · Cited by 19 · see citing works
Self-citationThe Role of AI-Driven De Novo Protein Design in the Exploration of the Protein Functional Universe
Zhang, G., Liu, C., Lu, J., Zhang, S., & Zhu, L. (2025) · Biology · doi:10.3390/biology14091268 · Cited by 20 · see citing works
“OpenFold [71] provides an open-source reimplemen-tation of AF2 with a comparable accuracy but a faster speed and lower memory usage, while SPIRED [72] can address high-throughput demands by delivering significant speed improvements and reduced computational costs while achieving a performance…”
Self-citationAssessing protein model quality based on deep graph coupled networks using protein language model
(2023) · doi:10.1093/bib/bbad420 · Cited by 18 · see citing works
Self-citationPredicting TCR sequences for unseen antigen epitopes using structural and sequence features
Ji, H., Wang, X.-X., Zhang, Q., Zhang, C., & Zhang, H. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae210 · Cited by 17 · see citing works
“OpenFold [19] is an open-source PyTorch reimplementation of AlphaFold2 [16] trained from scratch.”
Zhang, Z., Qian, H., Xu, R., Duan, Q., Wang, Y., Fu, T., Wu, X., Sha, C., Wu, Q., Wang, Y., et al. (2025) · Journal of the American Chemical Society · doi:10.1021/jacs.5c11823 · Cited by 17 · see citing works
Self-citation(2024) · Nature Methods · doi:10.1038/s41592-024-02391-7 · Cited by 17 · see citing works
Who will make AlphaFold3 open source? Scientists race to crack AI model
Callaway, E. (2024) · Nature · doi:10.1038/d41586-024-01555-x · Cited by 17 · see citing works
- Authors
- Ahdritz, G., Bouatta, N., Floristean, C., Kadyan, S., Xia, Q., Gerecke, W., O’Donnell, T. J., Berenberg, D., Fisk, I., Zanichelli, N., Zhang, B., Nowaczynski, A., Wang, B., Stepniewska-Dziubinska, M. M., Zhang, S., Ojewole, A., Guney, M. E., Biderman, S., Watkins, A. M., Ra, S., Lorenzo, P. R., Nivon, L., Weitzner, B., Ban, Y.-E. A., Chen, S., Zhang, M., Li, C., Song, S. L., He, Y., Sorger, P. K., Mostaque, E., Zhang, Z., Bonneau, R., & AlQuraishi, M.
- Affiliations
- Harvard University (US)
- Columbia University (US)
- Icahn School of Medicine at Mount Sinai (US)
- Courant Institute of Mathematical Sciences (US)
- Flatiron Institute (US)
- University of Utah (US)
- Nvidia (United States) (US)
- Booz Allen Hamilton (United States) (US)
- Genentech (US)
- Arzeda (United States) (US)
- Rutgers, The State University of New Jersey (US)
- University of Illinois Urbana-Champaign (US)
- Microsoft (United States) (US)
- Corresponding author
- Bouatta, Nazim
- AlQuraishi, Mohammed
- ORCID
- Ahdritz, Gustaf
- Bouatta, Nazim
- Floristean, Christina
- Kadyan, Sachin
- Xia, Qinghui
- Gerecke, William
- O’Donnell, Timothy J.
- Berenberg, Daniel
- Fisk, Ian
- Zanichelli, Niccolò
- Zhang, Bo
- Nowaczynski, Arkadiusz
- Wang, Bei
- Stepniewska-Dziubinska, Marta M.
- Zhang, Shang
- Ojewole, Adegoke
- Biderman, Stella
- Watkins, Andrew M.
- Ra, Stephen
- Lorenzo, Pablo Ribalta
- Weitzner, Brian
- Ban, Yih-En Andrew
- Chen, Shiyang
- Zhang, Minjia
- Song, Shuaiwen Leon
- He, Yuxiong
- Sorger, Peter K.
- Zhang, Zhao
- Bonneau, Richard
- AlQuraishi, Mohammed
- Venue
- Nature Methods
- Published
- 2024-05-14
- Volume
- 21
- Issue
- 8
- Pages
- 1514-1524
- ISSN
- 1548-7091 (ISSN-L)
- 1548-7105
- PMC
- PMC11645889
- OpenAlex
- W4396895148
- Cited by
- 439
- References
- 75
- Citations by year
- 2026: 133 · 2025: 138 · 2024: 107 · 2023: 58 · 2022: 3
- FWCI
- 51.31
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Nature Methods (publishedVersion, cc-by)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion, cc-by)
- Digital Access to Scholarship at Harvard (DASH) (Harvard University) (acceptedVersion)
- Funders
- U.S. Department of Health & Human Services | NIH | National Institute of General Medical Sciences (R35GM150546)
- U.S. Department of Health & Human Services | NIH | National Cancer Institute (U54-CA225088)
- National Science Foundation (OAC-2106661, OAC-2112606)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Physical Sciences › Materials Science › Materials Chemistry › Enzyme Structure and Function
- Keywords
- AlphaFold2, protein structure prediction, protein folding, model generalization, secondary structure elements, learning mechanisms, hierarchical folding
- MeSH terms
- Algorithms, Models, Molecular, Protein Conformation, Proteins, Software, Protein Structure, Secondary, Protein Folding, Computational Biology
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
