The DOI matches its Crossref record; 77 of the 77 references it cites were found in a registry or an index.
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.
Overview
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
Verification
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“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]
Experimental — accuracy not yet measured
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
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
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×
“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×
“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
Kabsch, W., & Sander, C. (1983) · Biopolymers · doi:10.1002/bip.360221211 · Cited by 15,320 · cited 9×
“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
Scaling 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 · doi:10.48550/arxiv.2001.08361 · Cited by 9,550
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×
“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×
“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
Cited by
Sorted by citations
Items 101–200 of 440
Accurate predictions of conformational ensembles of disordered proteins with STARLINGPreprint
Novak, B., Lotthammer, J. M., Emenecker, R. J., & Holehouse, A. S. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.02.14.638373 · Cited by 17
ProtMamba: a homology-aware but alignment-free protein state space model
Sgarbossa, D., Malbranke, C., & Bitbol, A.-F. (2025) · Bioinformatics · doi:10.1093/bioinformatics/btaf348 · Cited by 27
How artificial intelligence is reengineering protein engineering
Listgarten, J., & Jiang, H. (2026) · Science · doi:10.1126/science.aec8444 · Cited by 20
Leveraging Cross-Linking Mass Spectrometry for Modeling Antibody–Antigen Complexes
Ianni, A. D., Ianni, A. D., Cowan, K. J., Barbero, L. M., & Sirtori, F. R. (2024) · Journal of Proteome Research · doi:10.1021/acs.jproteome.3c00816 · Cited by 16
Structural prediction of GluN3 NMDA receptors
Liu, Y., Shao, D., Lou, S., & Kou, Z. (2024) · Frontiers in Physiology · doi:10.3389/fphys.2024.1446459 · Cited by 16
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
“We plan to incorporate additional modeling backends, including OpenFold.”
Probing the modulation of enzyme kinetics by multi-temperature, time-resolved serial crystallography
Schulz, E. C., Prester, A., von Stetten, D., Gore, G., Hatton, C. E., Bartels, K., Leimkohl, J.-P., Schikora, H., Ginn, H. M., Tellkamp, F., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-61631-2 · Cited by 15
Artificial intelligence in drug discovery — what it is, where we stand and the path forward
Bender, A., Thomas, M. C., Scannell, J. W., Shaywitz, D. A., Ghiandoni, G. M., Greener, J. G., Pruteanu, L.-L., Jacobson, R. D., Handa, K., Hirano, M., et al. (2026) · Nature Reviews Drug Discovery · doi:10.1038/s41573-026-01496-2 · Cited by 23
“Open-source models in the area of co-folding have largely caught up with closed-source developments, in particular in the context of the OpenFold consortium 80 .”
ExplainableFold: Understanding AlphaFold Prediction with Explainable AI
Tan, J., & Zhang, Y. (2023) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3580305.3599337 · Cited by 20
“More specifically, we use the OpenFold [2] implementation, and load the official pre-trained AlphaFold parameters 2 .”
MethodConditioned Protein Structure Prediction
(2024) · doi:10.1103/prxlife.2.043001 · Cited by 16
Mejı́a-Rodrı́guez, D., Kim, H. S., Sadler, N., Li, X., Bohutskyi, P., Valiev, M., Qian, W., & Cheung, M. S. (2023) · Protein Science · doi:10.1002/pro.4822 · Cited by 14
“The only requirement to run AlphaFold or OpenFold is to have the database files available.”
BackgroundMethodInfluentialAccurate Conformation Sampling via Protein Structural Diffusion
Fan, J., Li, Z., Alcaide, E., Ke, G., Huang, H., & E, W. (2024) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.4c00928 · Cited by 17
Challenging AlphaFold in predicting proteins with large-scale allosteric transitions
Perkins-Jechow, B. H., Ahualli, J. P. I., Nhu, H. T., Omidi, A., Li, C., Holguin-Cruz, J. A., Cho, D., Na, D., Malhis, N., Bui, J. M., et al. (2025) · Communications Chemistry · doi:10.1038/s42004-025-01763-0 · Cited by 14
“Achieving accurate, generalizable predictions of alternative conformations will ultimately require training on a broader and more phylogenetically diverse structural dataset 5,14 .”
Self-citationPredictive modeling and cryo-EM: A synergistic approach to modeling macromolecular structure
Corum, M. R., Venkannagari, H., Hryc, C. F., & Baker, M. L. (2024) · Biophysical Journal · doi:10.1016/j.bpj.2024.01.021 · Cited by 15
Integrating Large-Scale Protein Structure Prediction into Human Genetics Research
Marrero, M. C., Jänes, J., Baptista, D., & Beltrão, P. (2024) · Annual Review of Genomics and Human Genetics · doi:10.1146/annurev-genom-120622-020615 · Cited by 13
On knowing a gene: A distributional hypothesis of gene function
Kwon, J. J., Pan, J., Gonzalez, G., Hahn, W. C., & Žitnik, M. (2024) · Cell Systems · doi:10.1016/j.cels.2024.04.008 · Cited by 14
Machine learning in biological physics: From biomolecular prediction to design
Martin, J., Mateos, M., Onuchic, J. N., Coluzza, I., & Morcos, F. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2311807121 · Cited by 14
The human touch: Utilizing AlphaFold 3 to analyze structures of endogenous metabolons
Träger, T. K., Tüting, C., & Kastritis, P. L. (2024) · Structure · doi:10.1016/j.str.2024.08.018 · Cited by 14
Barlow Twins deep neural network for advanced 1D drug–target interaction prediction
Schuh, M. G., Boldini, D., Bohne, A. I., & Sieber, S. A. (2025) · Journal of Cheminformatics · doi:10.1186/s13321-025-00952-2 · Cited by 13
AlphaFold2's training set powers its predictions of some fold‐switched conformations
Schafer, J. W., & Porter, L. L. (2025) · Protein Science · doi:10.1002/pro.70105 · Cited by 12
Leveraging AI Advances and Online Tools for Structure‐Based Variant Analysis
Guzmán‐Vega, F. J., González‐Álvarez, A. C., Peña‐Guerra, K. A., Cardona‐Londoño, K. J., & Arold, S. T. (2023) · Current Protocols · doi:10.1002/cpz1.857 · Cited by 12
Dabrowski-Tumanski, P., & Stasiak, A. (2023) · Molecules · doi:10.3390/molecules28227462 · Cited by 12
“This, in turn, fostered numerous modifications and implementations [6–14].”
BackgroundComparing Power Signatures of HPC Workloads: Machine Learning vs Simulation
Govind, A., Bhalachandra, S., Zhao, Z., Rrapaj, E., Austin, B. M., & Nam, H. A. (2023) · International Conference for High Performance Computing, Networking, Storage and Analysis (SC) · doi:10.1145/3624062.3624274 · Cited by 12
“OpenFold is a PyTorch implementation of Google AlphaFold [6], which can accurately predict 3D models of protein structures [7].”
MethodFAPM: functional annotation of proteins using multimodal models beyond structural modeling
Xiang, W., Xiong, Z., Chen, H., Xiong, J., Zhang, W., Fu, Z., Zheng, M., Liu, B., & Shi, Q. (2024) · Bioinformatics · doi:10.1093/bioinformatics/btae680 · Cited by 12
“These trained models are then applied to specific tasks of predicting protein structure and properties using fine-tuning and transfer learning, leading to improved performance in biomedical research, drug design, and other related fields (Ahdritz et al. 2024, Clifford et al. 2022, Qiu et al. 2023).”
The physics-AI dialogue in drug design
Vargas-Rosales, P. A., & Caflisch, A. (2025) · RSC Medicinal Chemistry · doi:10.1039/d4md00869c · Cited by 15
Protein design accelerates the development and application of optogenetic tools
Zhu, L., Wang, Y., Wu, X., Wu, G., Zhang, G., Liu, C., & Zhang, S. (2025) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2025.02.014 · Cited by 13
Self-citationA new age in structural S-layer biology: Experimental and in silico milestones
Grill-Walcher, S., & Schäffer, C. (2025) · Journal of Biological Chemistry · doi:10.1016/j.jbc.2025.110205 · Cited by 12
Structural biology at the scale of proteomes
Bouatta, N., & AlQuraishi, M. (2023) · Nature Structural & Molecular Biology · doi:10.1038/s41594-023-00924-w · Cited by 12
Self-citationHetmann, M., Parigger, L., Sirelkhatim, H., Stern, A. C., Krassnigg, A., Gruber, K. F., Steinkellner, G., Ruau, D., & Gruber, C. C. (2024) · Scientific Data · doi:10.1038/s41597-024-03403-z · Cited by 11
PXDesign: Fast, Modular, and Accurate De Novo Design of Protein BindersPreprint
Protenix Team, Ren, M., Sun, J., Guan, J., Liu, C., Gong, C., Wang, Y. X., Wang, L., Cai, Q., Ma, W., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.08.15.670450 · Cited by 14
“Likewise, confidence scores such as pLDDT or interface pTM from the AlphaFold series and related variants [1, 2, 14, 19, 27, 30, 32, 32, 40, 47, 56, 57] are widely used for filtering, but important questions remain: How accurate and generalizable are these scores?”
Enabling Large Dynamic Neural Network Training with Learning-based Memory Management
Ren, J., Xu, D., Yang, S., Zhao, J., Li, Z., Navasca, C., Wang, C., Xu, H., & Li, D. (2024) · IEEE International Symposium on High-Performance Computer Architecture (HPCA) · doi:10.1109/hpca57654.2024.00066 · Cited by 14
DataSAIL: Data Splitting Against Information LeakagePreprint
Joeres, R., Blumenthal, D. B., & Kalinina, O. V. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.11.15.566305 · Cited by 11
“The most prominent examples are DeepMind’s AlphaFold[1] for predicting protein three-dimensional structure, introducing similar models, such as OmegaFold[2] and OpenFold[3].”
MethodFrom sequence to mechanobiology? Promises and challenges for AlphaFold 3
Zonta, F., & Pantano, S. (2024) · Mechanobiology in Medicine · doi:10.1016/j.mbm.2024.100083 · Cited by 11
“Another signi fi cant development is OpenFold, 13 an open-source implementation of AlphaFold that fosters transparency and enables further innovations.”
Xia, W., Liu, S., Huang, L., Wu, H., Bai, T., Wang, S., Hao, F., Wang, X. D., & Wang, W. (2024) · Molecular Catalysis · doi:10.1016/j.mcat.2024.114439 · Cited by 10
AI-accelerated therapeutic antibody development: practical insights
Santuari, L., Salvy, M. B., Xénarios, I., & Arpat, A. B. (2024) · Frontiers in Drug Discovery · doi:10.3389/fddsv.2024.1447867 · Cited by 10
Discovery, design, and engineering of enzymes based on molecular retrobiosynthesis
Chen, A., Peng, X., Shen, T., Zheng, L., Wu, D., & Wang, S. (2025) · mLife · doi:10.1002/mlf2.70009 · Cited by 12
“Currently, large language models (LLMs) based on the Transformer architecture have played a crucial role in protein sequence modeling 26 – 28 , protein structure prediction 29 – 31 , protein – ligand interactions 32,33 , and protein structure/sequence design 16,34 .”
Enhancing protein structure predictions: DeepSHAP as a tool for understanding AlphaFold2
Sibli, S. A., Panagiotou, V. P., & Makris, C. H. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127853 · Cited by 11
Luo, Y., & Cai, J. (2025) · Current Proteomics · doi:10.2174/0115701646374018250227075247 · Cited by 14
The Phenix‐AlphaFold webservice: Enabling AlphaFold predictions for use in Phenix
Poon, B. K., Terwilliger, T. C., & Adams, P. D. (2024) · Protein Science · doi:10.1002/pro.4992 · Cited by 9
“TensorFlow is the underlying framework used for AlphaFold and PyTorch is used for Open-Fold (Ahdritz et al., 2023), a structure prediction software package that is similar to AlphaFold.”
MethodHidden Structural States of Proteins Revealed by Conformer Selection with AlphaFold-NMRPreprint
Huang, Y. J., Ramelot, T. A., Spaman, L. E., Kobayashi, N., & Montelione, G. T. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.06.26.600902 · Cited by 9
Distilling Structural Representations into Protein Sequence ModelsPreprint
Ouyang-Zhang, J., Gong, C., Zhao, Y., Krähenbühl, P., Klivans, A. R., & Diaz, D. J. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.11.08.622579 · Cited by 12
Advancing Protein Ensemble Predictions Across the Order–Disorder ContinuumPreprint
Invernizzi, M., Bottaro, S., Streit, J. O., Trentini, B., Venanzi, N. A. E., Reidenbach, D., Lee, Y., Dallago, C., Sirelkhatim, H., Jing, B., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.10.18.680935 · Cited by 13
“Structure Module: The refined representations are passed to an invariant point attention (IPA)- based structure[111] module that predicts backbone frames, side-chain conformations, and atomic coordinates.”
AlphaFold3 for Structure-guided Ligand DiscoveryPreprint
Menon, K. M., Davasam, A., Chen, G., Bryant, C., Deng, Z., Lam, B., Yang, C., Barcelos, D., Liu, F., Alon, A., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2025.12.04.692352 · Cited by 15
Fadini, A., Li, M., McCoy, A. J., Banjara, S., Okumura, H., Napier, E., Fontana, P., Khan, A. R., Jovine, L., Terwilliger, T. C., et al. (2026) · Nature Methods · doi:10.1038/s41592-026-03047-4 · Cited by 22
“) to maximize agreement between an OpenFold prediction ( x prediction ) and experimental data ( E e ).”
InfluentialSelf-citationAllostery Illuminated: Harnessing AI and Machine Learning for Drug Discovery
Blanco, M.-J., Buskes, M. J., Govindaraj, R. G., Ipsaro, J. J., Prescott-Roy, J., & Padyana, A. K. (2024) · ACS Medicinal Chemistry Letters · doi:10.1021/acsmedchemlett.4c00260 · Cited by 9
Majidian, S., Hadziahmetovic, A., Langschied, F., Pascarelli, S., Prieto-Baños, S., Rojas-Vargas, J., Arvestad, L., Cheema, J., Cosentino, S., Ebersberger, I., et al. (2025) · Journal of Molecular Evolution · doi:10.1007/s00239-025-10272-6 · Cited by 12
“This explosion of CSMs is the result of groundbreaking advances in the field of computational protein structure prediction, driven by tools like AlphaFold (Jumper et al. 2021), RoseTTAFold (Baek et al. 2021), and OpenFold (Ahdritz et al. 2024).”
Flöge, K., Udayakumar, S., Sommer, J., Piraud, M., Kesselheim, S., Fortuin, V., Günnemann, S., van der Weg, K. J., Gohlke, H., Merdivan, E., et al. (2025) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1013679 · Cited by 13
“Importantly, their moderate size allows for a seamless application to new downstream tasks when needed.”
Molecular determinants of acrylamide neurotoxicity through covalent docking
Mueller, N. P. F., Carloni, P., & Alfonso‐Prieto, M. (2023) · Frontiers in Pharmacology · doi:10.3389/fphar.2023.1125871 · Cited by 11
“However, recently developed machine learning-based protein structure prediction algorithms (Baek et al., 2021; Jumper et al., 2021; Ahdritz et al., 2022; Lin et al., 2022) could also be used to generate input protein structures.”
MethodBonds and bytes: The odyssey of structural biology
Hoff, S., Zinke, M. C., Izadi‐Pruneyre, N., & Bonomi, M. (2023) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2023.102746 · Cited by 10
“Tools such as AF2, RoseTTAFold [11], OmegaFold [12], and OpenFold [13], which enable quick and accurate protein”
MethodStability Oracle: A Structure-Based Graph-Transformer for Identifying Stabilizing MutationsPreprint
Diaz, D. J., Gong, C., Ouyang-Zhang, J., Loy, J. M., Wells, J., Yang, D. K., Ellington, A. D., Dimakis, A. G., & Klivans, A. R. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.05.15.540857 · Cited by 9
“Deep learning is currently revolutionizing many physical and biological disciplines[17, 18, 19, 20, 21], with AlphaFold2 leading the way as the scientific breakthrough of the year in 2021[22, 23] and sparking an entire wave of deep learning structure prediction tools[24, 25].”
BackgroundWang, Y., Wang, Z., Sadeh, G., Zancato, L., Achille, A., Karypis, G., & Rangwala, H. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.10.29.620988 · Cited by 12
“We then use the Louvain algorithm (De Meo et al., 2011) to detect 8 communities (corresponding to 8 species) in this subgraph.”
InfluentialExperimental evaluation of AI-driven protein design risks using safe biological proxiesPreprint
Ikonomova, S. P., Wittmann, B. J., Piorino, F., Ross, D. J., Schaffter, S. W., Vasilyeva, O. B., Horvitz, E., Diggans, J., Strychalski, E. A., Lin‐Gibson, S., et al. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.05.15.654077 · Cited by 15
Protein Modeling with DEER Spectroscopy
Tessmer, M. H., & Stoll, S. (2024) · Annual Review of Biophysics · doi:10.1146/annurev-biophys-030524-013431 · Cited by 11
Norrild, R. K., von Bülow, S., Zanchet, B., Halldorsson, E., Lindorff‐Larsen, K., Rogers, J. M., & Buell, A. K. (2026) · Nature Communications · doi:10.1038/s41467-026-74825-z · Cited by 9
“Such interactions can lead to percolation coupled to phase separation into dynamic and viscoelastic condensates 4 .”
Latham, A. P., Zhang, W., Tempkin, J. O. B., Otsuka, S., Ellenberg, J., & Sali, A. (2025) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2415674122 · Cited by 8
AlphaFold2 SLiM screen for LC3-LIR interactions in autophagy
Stuke, J. F. M., & Hummer, G. (2025) · Autophagy · doi:10.1080/15548627.2025.2493999 · Cited by 8
“Additionally, our approach can easily be adapted for other structure prediction methods, such as OpenFold [75] and RoseTTAFold2 [76], though we did not test how this will affect results.”
Mahmud, M. I., & Banerjee, T. (2026) · Briefings in Bioinformatics · doi:10.1093/bib/bbag229 · Cited by 12
“AlphaFold and OpenFold DL methods predict protein structures with near-experimental accuracy from amino acid sequences [6, 113].”
ScaleFold: Reducing AlphaFold Initial Training Time to 10 Hours
Zhu, F., Nowaczynski, A., Li, R., Xin, J., Song, Y., Marcinkiewicz, M., Eryilmaz, S. B., Yang, J., & Andersch, M. (2024) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1145/3649329.3657326 · Cited by 8
Self-citationProtein ligand structure prediction: From empirical to deep learning approaches
Zhou, G., & DiMaio, F. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.102998 · Cited by 9
Learning with Privileged Knowledge Distillation for Improved Peptide–Protein Docking
Zhang, Z., Verburgt, J. C., Kagaya, Y., Christoffer, C. W., & Kihara, D. (2025) · ACS Omega · doi:10.1021/acsomega.5c00967 · Cited by 8
Self-citationPourseif, M. M., Hosseini, S. A. B., Khoshraftar, S. H., & Omidi, Y. (2025) · Bioimpacts · doi:10.34172/bi.33072 · Cited by 8
“26 Nevertheless, recent advances in structure prediction and molecular modeling such as AlphaFold 3, 27 OpenFold, 28 and NeuralPlexer, 29 have begun to alleviate these constraints by achieving near–QM-level accuracy with markedly improved scalability, enabling broader in silico screening and…”
Phased high-quality genome of the gymnosperm Himalayan Yew assists in paclitaxel pathway exploration
Li, Z., Zong, H., Liu, X., Wang, X., Liu, S., Jiao, X., Chen, X., Wu, H., Liu, Z., Wang, Z., et al. (2025) · GigaScience · doi:10.1093/gigascience/giaf026 · Cited by 6
“Structur e pr epar ation: All ODD enzyme structur es in this study were modeled using OpenFold (OpenFold 2.0) [ 50 , 51 ] with default parameters .”
InfluentialBalasco, N., Esposito, L., & Vitagliano, L. (2025) · Biomolecules · doi:10.3390/biom15050674 · Cited by 6
“It is virtually impossible to summarize the impressive impact of AF and its successors in the life sciences [30,35,46–53].”
Biomolecular Interaction Prediction in the Pre‐ and Post‐ AlphaFold Era: The 8th CAPRI Evaluation
Lensink, M. F., RAOURAOUA, N., Brysbaert, G., Velankar, S. S., Wodak, S. J., & Bonvin, A. M. J. J. (2025) · Proteins Structure Function and Bioinformatics · doi:10.1002/prot.70018 · Cited by 9
“This catalyzed the development of many other AI-based prediction methods like for example, RoseTTafold [14], AF-multimer [15], OpenFold [16], and many others.”
Xu, J., Ameerjan, A. B. P., Hungerland, J., Langebrake, G., Ravnsborg, T., Jensen, O. N., Schmidt, J., Bartölke, R., Kasahara, T., Satish, B., et al. (2025) · Journal of The Royal Society Interface · doi:10.1098/rsif.2025.0176 · Cited by 7
ESMDynamic: Fast and accurate prediction of protein dynamic contact maps from single sequences
Kleiman, D. E., Feng, J., Xue, Z., & Shukla, D. (2026) · Nature Communications · doi:10.1038/s41467-026-76361-2 · Cited by 7
Mahtha, S. K., Sureshkumar, V., & Mohanty, D. (2026) · NAR Genomics and Bioinformatics · doi:10.1093/nargab/lqag002 · Cited by 8
“…several other machine learning/deep learning/AI-based protein structure prediction meth-ods have emerged, such as RoseTTAFold [ 12 ], ESMFold [ 13 ], OpenFold [ 14 ], and OmegaFold [ 15 ] that require only the protein sequences as input to obtain highly reliable 3D protein structures [ 16 , 17 ].”
Morehead, A., Liu, J., Neupane, P., & Cheng, J. (2026) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2026.103247 · Cited by 8
Harnessing Deep Learning Methods for Voltage-Gated Ion Channel Drug Discovery
Mateos, D. L., Harris, B. J., González, A. H., Narang, K., & Yarov-Yarovoy, V. M. (2024) · Physiology · doi:10.1152/physiol.00029.2024 · Cited by 9
Prediction of structural variation
Kalakoti, Y., Sanjeev, A., & Wallner, B. (2025) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2025.103003 · Cited by 8
Greenshields-Watson, A., Abanades, B., & Deane, C. M. (2024) · Frontiers in Immunology · doi:10.3389/fimmu.2024.1352703 · Cited by 7
Robust Prediction of Multiple Protein Conformations with Entropy GuidancePreprint
Wu, D. T., & Feng, L. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.04.26.650728 · Cited by 7
“We implemented our entropy guidance method on top of OpenFold (a reproduction of AlphaFold2 in PyTorch) 36 , following the procedure outlined in Algorithm 2.”
Deep-learning structure elucidation from single-mutant deep mutational scanning
(2025) · doi:10.1038/s41467-025-62261-4 · Cited by 10
“When preparing sequences for mutational scanning, Tsuboyama et al. trimmed N-and C-terminal amino acids with a low number of contacts predicted by AlphaFold2 46 .”
AlphaFold model quality self‐assessment improvement via deep graph learning
Verburgt, J., Zhang, Z., & Kihara, D. (2025) · Protein Science · doi:10.1002/pro.70274 · Cited by 8
“The first example (Figure 3a) is a 208-residue long α -helical protein.”
Self-citationTremmel, R., Honoré, A., Park, Y., Zhou, Y.-T., Xiao, M., & Lauschke, V. (2025) · Pharmacogenomics (London) · doi:10.1080/14622416.2025.2504863 · Cited by 6
Open and sustainable AI: challenges, opportunities and the road ahead in the life sciences
Farrell, G., Adamidi, E. S., Buono, R. A., Anton, M., Attafi, O. A., Capella-Gutiérrez, S., Capriotti, E., Castro, L. J., Cirillo, D., Crossman, L. C., et al. (2026) · Nature Methods · doi:10.1038/s41592-026-03037-6 · Cited by 12
“This impeded its full reuse potential, which necessitated a more open community reimplementation, OpenFold 30 , that shared all model-relevant Perspective https://doi.org/10.1038/s41592-026-03037-6 data and code to: facilitate a reproduction of an AI model to enhance understanding; and reproduce consistent model bench…”
InfluentialHarnessing viral strategies to reverse cognitive dysfunction through the integrated stress response
Reineke, L. C., Zhu, P., Dalwadi, U., Dooling, S. W., Liu, Y., Wang, I.-C., Young-Baird, S. K., Okoh, J. T., Kuncha, S. K., Zhou, H., et al. (2026) · Science · doi:10.1126/science.aea8782 · Cited by 7
Abriata, L. A. (2026) · Communications Biology · doi:10.1038/s42003-026-10112-3 · Cited by 8
Risk assessment of homologous variants of biotech trait proteins using a bridging approach
Wang, C., Calcaterra, J., Anderson, B., Rydel, T. J., Wang, R., Bertho, L., Saracco, S. A., Hodge-Bell, K., Burzio, L. A., White, T., et al. (2024) · GM crops & food · doi:10.1080/21645698.2024.2420412 · Cited by 6
“The folded protein structures for nsDMO, nmDMO, sbDMO, along with wild type DMO (WT DMO) were obtained using locally-installed versions of the artificial intelligence-driven computational protein folding programs AlphaFold, 38 OpenFold, 39 and ESMFold.”
Sidechain conditioning and modeling for full-atom protein sequence design with FAMPNNPreprint
Shuai, R. W., Widatalla, T., Huang, P., & Hie, B. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.02.13.637498 · Cited by 10
How AlphaFold and related models predict protein-peptide complex structuresPreprint
Guan, L., & Keating, A. (2025) · bioRxiv · doi:10.1101/2025.06.18.660495 · Cited by 7
Lessons from the Trenches on Reproducible Evaluation of Language Models
Biderman, S., Schoelkopf, H., Sutawika, L., Gao, L., Tow, J., Abbasi, B., Aji, A. F., Ammanamanchi, P. S., Black, S., Clive, J., et al. (2024) · arXiv · doi:10.48550/arxiv.2405.14782 · Cited by 9
“…GPT-3 (Brown et al., 2020), and ChatGPT, used as approximate stand-ins for shifts in how the community uses and therefore evaluates LMs. Common practice for evaluating autoregressive language models today diverges from the method described in the paper for all listed tasks except MMLU and MATH.”
Self-citationDay, E. H., & Lindert, S. (2025) · Journal of the American Society for Mass Spectrometry · doi:10.1021/jasms.5c00053 · Cited by 4
AlphaFold distillation for inverse protein design
Melnyk, I., Lozano, A. C., Das, P., & Chenthamarakshan, V. (2025) · Scientific Reports · doi:10.1038/s41598-025-00436-1 · Cited by 9
“Timings for AlphaFold and OpenFold 16 do not include MSA search times, which can range from minutes to hours.”
AI-based hardware and software tools in microscopy to boost research in immunology and virology
Morone, D., & D’Antuono, R. (2025) · Frontiers in Immunology · doi:10.3389/fimmu.2025.1610345 · Cited by 5
Swapna, G. V. T., Dube, N., Roth, M. J., & Montelione, G. T. (2025) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1013590 · Cited by 9
“While AlphaFold2/3 (AF2/3) [1], Evolutionary Scale Modeling (ESM) [2], and related machine-learning methods [3,4] can provide accurate structural models of proteins, for systems that adopt multiple conformational states, conventional AF2/3 and ESM calculations often deliver only one of the multiple…”
Čondić‐Jurkić, K., Alibay, I., Sherman, W., Tollefson, M. R., Walters, W. P., Baker, Z., Chong, L. T., Wei, J. N., Gray, J. J., Weitzner, B. D., et al. (2026) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.5c03137 · Cited by 6
Self-citationMolViewSpec: a Mol* extension for describing and sharing molecular visualizations
Midlik, A., Bittrich, S., Fleming, J. R., Nair, S. S., Velankar, S. S., Burley, S. K., Young, J. Y., Vallat, B. K., & Sehnal, D. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkaf370 · Cited by 5
“New generation AI-powered tools like AlphaFold [ 5 ], ESMFold [ 6 ], and OpenFold [ 7 ] provide access to millions of predicted structures.”
Artificial intelligence and first‐principle methods in protein redesign: A marriage of convenience?
Cianferoni, D., Vizarraga, D., Fernández‐Escamilla, A. M., Fita, I., Hamdani, R., Reche, R., Delgado, J., & Serrano, L. (2025) · Protein Science · doi:10.1002/pro.70210 · Cited by 4
Integrative modeling meets deep learning: Recent advances in modeling protein assemblies
Shor, B., & Schneidman‐Duhovny, D. (2024) · Current Opinion in Structural Biology · doi:10.1016/j.sbi.2024.102841 · Cited by 6
Wang, C., Zheng, M., Est, C. B., Lawal, R., Liang, W. G., Korasick, D. A., Rau, M., Saracco, S. A., Johnson, V. R., Wang, Y., et al. (2024) · PLoS ONE · doi:10.1371/journal.pone.0311049 · Cited by 5
Solving the protein folding problem…
Sleator, R. D. (2024) · FEBS Letters · doi:10.1002/1873-3468.15043 · Cited by 7
Random, de novo and conserved proteins: How structure and disorder predictors perform differentlyPreprint
Middendorf, L., & Eicholt, L. A. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.07.18.549582 · Cited by 5
Unmasking AlphaFold: integration of experiments and predictions in multimeric complexesPreprint
Mirabello, C., Wallner, B., Nystedt, B., Azinas, S., & Carroni, M. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.09.20.558579 · Cited by 5
“Other integrative approaches propose retraining versions of AlphaFold that are better at taking experiments into account: versions of OpenFold (11) and Uni-Fold (12) (retrainable implementations of AlphaFold) accept extra input data derived from crosslinking experiments to guide monomeric and multimeric predictions (1…”
MethodProtein Language Model Supervised Precise and Efficient Protein Backbone Design MethodPreprint
Zhang, B., Liu, K., Zheng, Z., Liu, Y., Mu, J., Wei, T., & Chen, H. (2023) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2023.10.26.564121 · Cited by 5
Self-citationRaisinghani, N., Alshahrani, M., Gupta, G., Tian, H., Xiao, S., Tao, P., & Verkhivker, G. M. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.02.15.580591 · Cited by 4
FoldToken3: Fold Structures Worth 256 Words or LessPreprint
Gao, Z., Tan, C., & Li, S. Z. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.07.08.602548 · Cited by 4
“For example, training well-known AlphaFold2 from scratch takes 128 TPUv3 cores for 11 days [15]; OpenFold takes 50000 GPU hours for training [2].”
MethodAlphaFold2’s training set powers its predictions of fold-switched conformationsPreprint
Schafer, J. W., & Porter, L. L. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.10.11.617857 · Cited by 5
“Although this explanation is possible, previous research shows that AF2 performs nearly as well with a training set of 10,000 structures as it does with its full set of over 100,000 (Ahdritz et al. 2024).”
Method
Metadata
- 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
- 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
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Structural Biology › 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
