AbstractAccurate particle picking of macromolecules of diverse types, shapes and sizes in cryo-electron tomography (cryo-ET) is critical for understanding the molecular architecture of biological systems in their native state. Current deep learning methods have shown potential for identifying macromolecules in tomograms; however, they are vulnerable to annotation noise in the training data and to imbalanced distributions of macromolecule species. Here, to address these limitations, we developed RobPicker, a meta-learning framework that effectively mitigates these issues by automatically learning to correct label errors and by placing greater emphasis on smaller and rarer macromolecule species. In evaluations across diverse cryo-ET datasets with noisy labels and imbalanced species distributions, RobPicker substantially outperforms state-of-the-art methods, particularly in identifying small and rare macromolecules. The robustness and efficiency of RobPicker led to improved resolution in 3D density maps of protein structures, enabling high-resolution cellular structural biology analysis.This is a preview of subscription content, access via your institutionAccess options Access through your institutionAccess Nature and 54 other Nature Portfolio journalsGet Nature+, our best-value online-access subscription34.99 € / 30 days cancel any timeLearn moreSubscribe to this journalReceive 12 print issues and online access269.00 € per year only 22.42 € per issueLearn moreBuy this articlePurchase on SpringerLinkInstant access to the full article PDF.39,95 €Prices may be subject to local taxes which are calculated during checkoutFig. 1: To address noisy labels and species imbalance, RobPicker uses BLO with DR and LC to enable robust macromolecule segmentation.Fig. 2: RobPicker improves particle picking across real cryo-ET datasets.Fig. 3: RobPicker improves particle picking on the CZII benchmark.Fig. 4: Subtomogram averaging using RobPicker’s picks, ground-truth annotations and PyTom’s picks.Fig. 5: RobPicker shows high accuracy and strong cross-organism generalizability in particle picking.Fig. 6: RobPicker outperformed DeepFinder on the SHREC 2019 dataset.Data availabilityThe datasets are available via Zenodo at https://doi.org/10.5281/zenodo.22238302 (ref. 94). Dataset D1 is publicly available in the EMPIAR database under accession number EMPIAR-11830. The corresponding annotations are available via GitHub at https://github.com/Chromatin-Structure-Rhythms-Lab/ChlamyAnnotations. D2 is available at EMPIAR-12534. D3 and D4 are available from the corresponding author upon reasonable request because these internally generated datasets are not currently hosted in a public repository. D5 is available at EMPIAR-10988. The M. pneumoniae tomogram dataset is available at EMPIAR-10499. The dataset of the SHREC 2019 challenge can be accessed at http://www2.projects.science.uu.nl/shrec/cryo-et/2019/. The CZII datasets are publicly available through the CryoET Data Portal and can be accessed using their corresponding dataset IDs: 10440, 10445 and 10446. Source data are provided with this paper.Code availabilityThe source code of RobPicker is available via GitHub at https://github.com/RobPicker/CryoET-RobPicker and via Zenodo at https://doi.org/10.5281/zenodo.22261457 (ref. 95).ReferencesBeck, M. & Baumeister, W. Cryo-electron tomography: can it reveal the molecular sociology of cells in atomic detail?. Trends Cell Biol. 26, 825–837 (2016).Article PubMed Google Scholar Mahamid, J. et al. Visualizing the molecular sociology at the HeLa cell nuclear periphery. Science 351, 969–972 (2016).Article CAS PubMed Google Scholar McCafferty, C. L. et al. Integrating cellular electron microscopy with multimodal data to explore biology across space and time. Cell 187, 563–584 (2024).Article CAS PubMed Google Scholar Nogales, E. & Mahamid, J. Bridging structural and cell biology with cryo-electron microscopy. Nature 628, 47–56 (2024).Article CAS PubMed PubMed Central Google Scholar Cai, S. et al. Cryo-ET reveals the macromolecular reorganization of S. pombe mitotic chromosomes in vivo. Proc. Natl Acad. Sci. 115, 10977–10982 (2018).Article CAS PubMed PubMed Central Google Scholar Huang, C.-Y. et al. In situ structure and dynamics of an alphacoronavirus spike protein by cryo-ET and cryo-EM. Nat. Commun. 13, 4877 (2022).Article CAS PubMed PubMed Central Google Scholar Pyle, E., Miller, E. A. & Zanetti, G. Cryo-electron tomography reveals how COPII assembles on cargo-containing membranes. Nat. Struct. Mol. Biol. 32, 513–519 (2025).Article CAS PubMed Google Scholar Lučić, V., Rigort, A. & Baumeister, W. Cryo-electron tomography: the challenge of doing structural biology in situ. J. Cell Biol. 202, 407–419 (2013).Article PubMed PubMed Central Google Scholar Berger, C. et al. Cryo-electron tomography on focused ion beam lamellae transforms structural cell biology. Nat. Methods 20, 499–511 (2023).Article CAS PubMed Google Scholar Eisenstein, F. et al. Parallel cryo electron tomography on in situ lamellae. Nat. Methods 20, 131–138 (2023).Article CAS PubMed Google Scholar Pyle, E. & Zanetti, G. Current data processing strategies for cryo-electron tomography and subtomogram averaging. Biochem. J. 478, 1827–1845 (2021).Article CAS PubMed PubMed Central Google Scholar Watson, A. J. I. & Bartesaghi, A. Advances in cryo-ET data processing: meeting the demands of visual proteomics. Curr. Opin. Struct. Biol. 87, 102861 (2024).Article CAS PubMed PubMed Central Google Scholar Chen, M. et al. A complete data processing workflow for cryo-ET and subtomogram averaging. Nat. Methods 16, 1161–1168 (2019).Article CAS PubMed PubMed Central Google Scholar Eisenstein, F., Fukuda, Y. & Danev, R. Smart parallel automated cryo-electron tomography. Nat. Methods 21, 1612–1615 (2024).Article CAS PubMed Google Scholar Iudin, A., Korir, P. K., Torres, J., Kleywegt, G. J. & Patwardhan, A. EMPIAR: a public archive for raw electron microscopy image data. Nat. Methods 13, 387–388 (2016).Article CAS PubMed Google Scholar Gao, S. et al. Macromolecules structural classification with a 3D dilated dense network in cryo-electron tomography. IEEE/ACM Trans. Comput. Biol. Bioinform. 19, 209–219 (2021).Article Google Scholar Zeng, Y. et al. Unsupervised domain alignment based open set structural recognition of macromolecules captured by cryo-electron tomography. In 2021 IEEE International Conference on Image Processing 106–110 (IEEE, 2021).Lamm, L. et al. MemBrain: a deep learning-aided pipeline for detection of membrane proteins in cryo-electron tomograms. Comput. Methods Programs Biomed. 224, 106990 (2022).Article PubMed Google Scholar Zeng, X. et al. High-throughput cryo-ET structural pattern mining by unsupervised deep iterative subtomogram clustering. Proc. Natl Acad. Sci. USA 120, e2213149120 (2023).Article CAS PubMed PubMed Central Google Scholar Moebel, E. et al. Deep learning improves macromolecule identification in 3D cellular cryo-electron tomograms. Nat. Methods 18, 1386–1394 (2021).Article CAS PubMed Google Scholar de Teresa-Trueba, I. et al. Convolutional networks for supervised mining of molecular patterns within cellular context. Nat. Methods 20, 284–294 (2023).Article PubMed PubMed Central Google Scholar Liu, G. et al. DeepETPicker: fast and accurate 3D particle picking for cryo-electron tomography using weakly supervised deep learning. Nat. Commun. 15, 2090 (2024).Article CAS PubMed PubMed Central Google Scholar Çiçek, Ö., Abdulkadir, A., Lienkamp, S. S., Brox, T. & Ronneberger, O. 3D U-Net: learning dense volumetric segmentation from sparse annotation. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2016: 19th International Conference, Proceedings, Part II (eds Ourselin, S. et al.) 424–432 (Springer, 2016).Moebel, E. & Kervrann, C. Towards unsupervised classification of macromolecular complexes in cryo electron tomography: challenges and opportunities. Comput. Methods Programs Biomed. 225, 107017 (2022).Article CAS PubMed Google Scholar Rice, G. et al. TomoTwin: generalized 3D localization of macromolecules in cryo-electron tomograms with structural data mining. Nat. Methods 20, 871–880 (2023).Article CAS PubMed PubMed Central Google Scholar Yan, R., Venkatakrishnan, S. V., Liu, J., Bouman, C. A. & Jiang, W. MBIR: a cryo-ET 3D reconstruction method that effectively minimizes missing wedge artifacts and restores missing information. J. Struct. Biol. 206, 183–192 (2019).Article PubMed PubMed Central Google Scholar Turk, M. & Baumeister, W. The promise and the challenges of cryo-electron tomography. FEBS Lett. 594, 3243–3261 (2020).Article CAS PubMed Google Scholar Bepler, T., Kelley, K., Noble, A. J. & Berger, B. Topaz-Denoise: general deep denoising models for cryoEM and cryoET. Nat. Commun. 11, 5208 (2020).Article CAS PubMed PubMed Central Google Scholar Liu, Y.-T. et al. Isotropic reconstruction for electron tomography with deep learning. Nat. Commun. 13, 6482 (2022).Article CAS PubMed PubMed Central Google Scholar Zhang, H. et al. A method for restoring signals and revealing individual macromolecule states in cryo-ET, REST. Nat. Commun. 14, 2937 (2023).Article CAS PubMed PubMed Central Google Scholar Wiedemann, S. & Heckel, R. A deep learning method for simultaneous denoising and missing wedge reconstruction in cryogenic electron tomography. Nat. Commun. 15, 8255 (2024).Article CAS PubMed PubMed Central Google Scholar Chen, M. et al. Convolutional neural networks for automated annotation of cellular cryo-electron tomograms. Nat. Methods 14, 983–985 (2017).Article CAS PubMed PubMed Central Google Scholar Böhm, J. et al. Toward detecting and identifying macromolecules in a cellular context: template matching applied to electron tomograms. Proc. Natl Acad. Sci. USA 97, 14245–14250 (2000).Article PubMed PubMed Central Google Scholar Wagner, T. & Raunser, S. The evolution of SPHIRE-crYOLO particle picking and its application in automated cryo-EM processing workflows. Commun. Biol. 3, 61 (2020).Article PubMed PubMed Central Google Scholar Cruz-León, S. et al. High-confidence 3D template matching for cryo-electron tomography. Biophys. J. 123, 183a (2024).Article Google Scholar Ermel, U. et al. A data portal for providing standardized annotations for cryo-electron tomography. Nat. Methods 21, 2200–2202 (2024).Huang, Q., Zhou, Y. & Bartesaghi, A. MiLoPYP: self-supervised molecular pattern mining and particle localization in situ. Nat. Methods 21, 1863–1872 (2024).Article CAS PubMed PubMed Central Google Scholar Stojanovska, F. et al. CryoSiam: self-supervised representation learning for automated analysis of cryo-electron tomograms. Preprint at bioRxiv https://doi.org/10.1101/2025.11.11.687379. (2025).Bandyopadhyay, H. et al. Cryo-shift: reducing domain shift in cryo-electron subtomograms with unsupervised domain adaptation and randomization. Bioinformatics 38, 977–984 (2022).Article CAS PubMed PubMed Central Google Scholar Harastani, M., Patra, G., Kervrann, C. & Eltsov, M. Template learning: deep learning with domain randomization for particle picking in cryo-electron tomography. Nat. Commun. 16, 8833 (2025).Article CAS PubMed PubMed Central Google Scholar Luo, Z., Zeng, X., Bao, Z., & Xu, M., Deep learning-based strategy for macromolecules classification with imbalanced data from cellular electron cryotomography. In 2019 International Joint Conference on Neural Networks 1–8 (IEEE, 2019).Zhao, R. et al. Rethinking Dice Loss for medical image segmentation. In 2020 IEEE International Conference on Data Mining 851–860 (IEEE, 2020).Yeung, M., Sala, E., Schönlieb, C.-B. & Rundo, L. Unified focal loss: generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation. Comput. Med. Imaging Graph. 95, 102026 (2022).Article PubMed Google Scholar Li, Z., Kamnitsas, K. & Glocker, B. Analyzing overfitting under class imbalance in neural networks for image segmentation. IEEE Trans. Med. Imaging 40, 1065–1077 (2020).Article Google Scholar Hospedales, T., Antoniou, A., Micaelli, P. & Storkey, A. Meta-learning in neural networks: a survey. IEEE Trans. Pattern Anal. Mach. Intell. 44, 5149–5169 (2021).Google Scholar Harrington, K. et al. CZII - CryoET object identification. Kaggle https://kaggle.com/competitions/czii-cryo-et-object-identification (2024).Zheng, G., Awadallah, A. H. & Dumais, S. Meta label correction for noisy label learning. In Proc. AAAI Conference on Artificial Intelligence Vol. 35, 11053–11061 (AAAI, 2021).Gubins, I. et al. SHREC 2020: classification in cryo-electron tomograms. Comput. Graph. 91, 279–289 (2020).Article Google Scholar Shu, J. et al. Meta-weight-net: learning an explicit mapping for sample weighting. In Advances in Neural Information Processing Systems Vol. 32 (eds Wallach, H. et al.) (Curran Associates, 2019).Chaillet, M. L., Roet, S., Veltkamp, R. C. & Förster, F. pytom-match-pick: a tophat-transform constraint for automated classification in template matching. J. Struct. Biol. 11, 100125 (2025).CAS Google Scholar Burt, A. et al. An image processing pipeline for electron cryo-tomography in RELION-5. FEBS Open Bio. 14, 1788–1804 (2024).Article CAS PubMed PubMed Central Google Scholar Tegunov, D., Xue, L., Dienemann, C., Cramer, P. & Mahamid, J. Multi-particle cryo-EM refinement with M visualizes ribosome–antibiotic complex at 3.5 Å in cells. Nat. Methods 18, 186–193 (2021).Article CAS PubMed PubMed Central Google Scholar Dice, L. R. Measures of the amount of ecologic association between species. Ecology 26, 297–302 (1945).Article Google Scholar Liu, H., Simonyan, K. & Yang, Y. DARTS: Differentiable Architecture Search. In International Conference on Learning Representations (OpenReview.net, 2019).Choe, S. K., Neiswanger, W., Xie, P. & Xing, E. Betty: an automatic differentiation library for multilevel optimization. In The Eleventh International Conference on Learning Representations (OpenReview.net, 2023).Paszke, A. et al. PyTorch: an imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems Vol. 32 (eds Wallach, H. et al.) (Curran Associates, 2019).Ioffe, S. & Szegedy, C. Batch normalization: accelerating deep network training by reducing internal covariate shift. In International Conference on Machine Learning (eds Bach, F. & Blei, D.) 448–456 (PMLR, 2015).Nair, V. & Hinton, G. E. Rectified linear units improve restricted boltzmann machines. In Proc. 27th International Conference on Machine Learning (ICML-10) (eds Fürnkranz, J. & Joachims, T.) 807–814 (Omnipress, 2010).Henkel, C., & Khvedchenya, E., CZII - 1st place solution. Kaggle https://www.kaggle.com/competitions/czii-cryo-et-object-identification/writeups/daddies-1st-place-solution-segmentation-with-partl (accessed 17 November 2025).He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 770–778 (IEEE, 2016).Zhang, H., Cisse, M., Dauphin, Y. N. & Lopez-Paz, D. mixup: beyond empirical risk minimization. In International Conference on Learning Representations (OpenReview.net, 2018).Ester, M. et al. A density-based algorithm for discovering clusters in large spatial databases with noise. In KDD Vol. 96 (eds Simoudis, E. et al.) 226–231 (AAAI, 1996).Kremer, J. R., Mastronarde, D. N. & McIntosh, J. R. Computer visualization of three-dimensional image data using imod. J. Struct. Biol. 116, 71–76 (1996).Article CAS PubMed Google Scholar Meng, E. C. et al. UCSF ChimeraX: tools for structure building and analysis. Protein Sci. 32, e4792 (2023).Article CAS PubMed PubMed Central Google Scholar Kelley, R. et al. Toward community-driven visual proteomics with large-scale cryo-electron tomography of Chlamydomonas reinhardtii. Mol. Cell 86, 213–230 https://doi.org/10.1016/j.molcel.2025.11.029 (2026).Article CAS Google Scholar Chang, Y.-T. et al. Cytoplasmic ribosomes on mitochondria alter the local membrane environment for protein import. J. Cell Biol. 224, e202407110 (2025).Article CAS PubMed PubMed Central Google Scholar Tegunov, D. & Cramer, P. Real-time cryo-electron microscopy data preprocessing with warp. Nat. Methods 16, 1146–1152 (2019).Article CAS PubMed PubMed Central Google Scholar Wagner, F. R. et al. Preparing samples from whole cells using focused-ion-beam milling for cryo-electron tomography. Nat. Protoc. 15, 2041–2070 (2020).Article CAS PubMed PubMed Central Google Scholar Zivanov, J. et al. New tools for automated high-resolution cryo-EM structure determination in RELION-3. eLife 7, e42166 (2018).Article PubMed PubMed Central Google Scholar Tang, G. et al. EMAN2: an extensible image processing suite for electron microscopy. J. Struct. Biol. 157, 38–46 (2007).Article CAS PubMed Google Scholar Peck, A. et al. Lessons learned from a Kaggle challenge for particle picking in cryo-electron tomography. Nat. Methods 23, 1908–1916 (2026).Article CAS PubMed PubMed Central Google Scholar Gubins, I. et al. Classification in cryo-electron tomograms. Eurographics Workshop on 3D Object Retrieval https://doi.org/10.2312/3dor.20191061 (2019).Peck, A. et al. A realistic phantom dataset for benchmarking cryo-ET data annotation. Nat. Methods 22, 1819–1823 https://doi.org/10.1038/s41592-025-02800-5 (2025).Luo, Z., & Wang, S., CZII - 2nd Place Solution. Kaggle https://www.kaggle.com/competitions/czii-cryo-et-object-identification/writeups/luoziqian-lion-2nd-place-solution (accessed 17 November 2025).Ouertani, Y., Ouertani, E., Mehdi, W. & Souid, A., CZII - 5th Place Solution. Kaggle https://www.kaggle.com/competitions/czii-cryo-et-object-identification/writeups/youssef-ouertani-5th-place-solution (accessed 17 November 2025).Lin, T.-Y., Goyal, P., Girshick, R., He, K. & Dollár, P. Focal loss for dense object detection. In Proc. IEEE International Conference on Computer Vision 2980–2988 (IEEE, 2017).Cianfrocco, M. A., Lahiri, I., DiMaio, F. & Leschziner, A. E. cryoEM-cloud-tools: a software platform to deploy and manage cryo-EM jobs in the cloud. J. Struct. Biol. 203, 230–235 (2018).Article PubMed PubMed Central Google Scholar Adaixo, R. et al. I. Cryo-EM structure of native human thyroglobulin. Nat. Commun. 13, 61 (2022).Article CAS PubMed PubMed Central Google Scholar Amari, S.-I. Backpropagation and stochastic gradient descent method. Neurocomputing 5, 185–196 (1993).Article Google Scholar Ketkar, N. & Ketkar, N. in Deep learning with Python: A Hands-On Introduction 113–132 (Apress, 2017).Cryo et 1st place solution. Kaggle https://www.kaggle.com/code/christofhenkel/cryo-et-1st-place-solution?scriptVersionId=223259615 (accessed 17 November 2025).7-model-6TTA-T4x2-patch-norm-v2. Kaggle https://www.kaggle.com/code/luoziqian/7-model-6tta-t4x2-patch-norm-v2/notebook?scriptVersionId=220452693 (accessed 17 November 2025).inference_script. Kaggle https://www.kaggle.com/code/youssefouertani/inference-script (accessed 17 November 2025).Myronenko, A. 3D MRI brain tumor segmentation using autoencoder regularization. In Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries (eds Crimi, A. et al.) 311–320 (Springer, 2019)Isensee, F., Jaeger, P. F., Kohl, S. A. A., Petersen, J. & Maier-Hein, K. H. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat. Methods 18, 203–211 (2021).Article CAS PubMed Google Scholar Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization. In Third International Conference on Learning Representations (eds Bengio, Y. & LeCun, Y.) (2015).Çiçek, Ö., Abdulkadir, A., Lienkamp, S. S., Brox, T. & Ronneberger, O. 3D U-Net: learning dense volumetric segmentation from sparse annotation. In International Conference on Medical Image Computing and Computer-Assisted Intervention (eds Ourselin, S. et al.) (Springer, 2016).Chen, H., Dou, Q., Yu, L., Qin, J. & Heng, P.-A. VoxResNet: deep voxelwise residual networks for volumetric brain segmentation. NeuroImage 170, 446–455 (2018).Article PubMed Google Scholar Li, Y. et al. Whole brain segmentation with full volume neural network. Comput. Med. Imaging Graph. 93, 101991 (2021).Article PubMed Google Scholar Gibson, E. et al. Automatic multi-organ segmentation on abdominal CT with dense V-networks. IEEE Trans. Med. Imaging 37, 1822–1834 (2018).Article PubMed PubMed Central Google Scholar Zhou, Z., Siddiquee, M. M. R., Tajbakhsh, N. & Liang, J. UNet++: a nested U-net architecture for medical image segmentation. IEEE Trans. Med. Imaging 39, 1856–1867 (2020).Article PubMed Google Scholar Salehi, S. S. M., Erdogmus, D. & Gholipour, A. Tversky loss function for image segmentation using 3D fully convolutional deep networks. In International Workshop on Machine Learning in Medical Imaging (eds Wang, Q. et al.) 379–387 (Springer, 2017).Loshchilov, I. & Hutter, F. Decoupled weight decay regularization. In International Conference on Learning Representations (OpenReview.net, 2019).Liang, Y. et al. Datasets associated with RobPicker: a meta learning framework for robust identification of macromolecules in cryo-electron tomograms. Zenodo https://doi.org/10.5281/zenodo.22238302 (2026).Liang, Y. et al. Source code associated with RobPicker. Zenodo https://doi.org/10.5281/zenodo.22261457 (2026).Download referencesAcknowledgementsWe thank U. Ermel, A. Peck, K. Zhao, J. Hutchings, J. Schwartz and M. Paraan for providing helpful clarification about the CZII dataset and the CZII CryoET Object Identification Challenge. We thank F. Stojanovska, J. Zaugg and J. Mahamid for providing helpful clarification about the dataset D5. We thank J.-H. Schäfer for providing help and insightful discussion about the refinement and the FSC calculations.FundingP.X. acknowledges funding support from NSF IIS2405974, NSF IIS2339216 and NIH R35GM157217. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.Author informationAuthor notesThese authors contributed equally: Youwei Liang, Ramtin Hosseini.Authors and AffiliationsDepartment of Electrical and Computer Engineering, University of California San Diego, La Jolla, CA, USAYouwei Liang, Ramtin Hosseini, Ruiyi Zhang, Li Zhang, Han Guo & Pengtao XieDepartment of Integrative Structural and Computational Biology, The Scripps Research Institute, La Jolla, CA, USAHamidreza Rahmani, James Williamson & Danielle A. GrotjahnSchool of Biological Sciences, University of California San Diego, La Jolla, CA, USADigvijay Singh, Joan Lee & Elizabeth VillaDepartment of Computer Science, Carnegie Mellon University, Pittsburgh, PA, USASang ChoeComputational Biology Department, Carnegie Mellon University, Pittsburgh, PA, USAMin XuDepartment of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, IsraelEran SegalDepartment of Molecular Cell Biology, Weizmann Institute of Science, Rehovot, IsraelEran SegalDepartment of Biomedical Data Science, Stanford University, Stanford, CA, USAJames ZouDepartment of Computer Science, Stanford University, Stanford, CA, USAJames ZouHoward Hughes Medical Institute, University of California San Diego, La Jolla, CA, USAElizabeth VillaDepartment of Medicine, University of California San Diego, La Jolla, CA, USAPengtao XieAuthorsYouwei LiangView author publicationsSearch author on:PubMed Google ScholarRamtin HosseiniView author publicationsSearch author on:PubMed Google ScholarHamidreza RahmaniView author publicationsSearch author on:PubMed Google ScholarDigvijay SinghView author publicationsSearch author on:PubMed Google ScholarRuiyi ZhangView author publicationsSearch author on:PubMed Google ScholarLi ZhangView author publicationsSearch author on:PubMed Google ScholarHan GuoView author publicationsSearch author on:PubMed Google ScholarSang ChoeView author publicationsSearch author on:PubMed Google ScholarJoan LeeView author publicationsSearch author on:PubMed Google ScholarMin XuView author publicationsSearch author on:PubMed Google ScholarEran SegalView author publicationsSearch author on:PubMed Google ScholarJames ZouView author publicationsSearch author on:PubMed Google ScholarJames WilliamsonView author publicationsSearch author on:PubMed Google ScholarElizabeth VillaView author publicationsSearch author on:PubMed Google ScholarDanielle A. GrotjahnView author publicationsSearch author on:PubMed Google ScholarPengtao XieView author publicationsSearch author on:PubMed Google ScholarContributionsR.H. and P.X. proposed the initial concept of using LC and DR in training cryo-ET particle networks. R.H. and Y.L. implemented the code for RobPicker, and carried out training and prediction with the RobPicker models. Y.L., R.H., H.R. and D.S. performed analysis and evaluation of the cryo-ET picking results. H.R. and D.S. carried out subtomogram averaging and structural analysis. S.C. implemented the BLO algorithm. R.Z., L.Z. and H.G. carried out training and prediction with the cryo-ET picking baseline methods in the comparison with RobPicker. H.R., D.S., J.L., J.W., E.V. and D.A.G. acquired cryo-electron tomograms and co-produced ground-truth annotations. Y.L., R.H., H.R., D.S., R.Z., L.Z. and H.G. produced the figures with input from all authors. Y.L., R.H., H.R., D.S., R.Z., L.Z., H.G., J.Z., E.S., M.X., E.V., D.A.G. and P.X. wrote the paper with input from all authors.Corresponding authorCorrespondence to Pengtao Xie.Ethics declarationsCompeting interestsThe authors declare no competing interests.Peer reviewPeer review informationNature Methods thanks Kyle Harrington, William Wan and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editor: Arunima Singh, in collaboration with the Nature Methods team.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Extended dataExtended Data Fig. 1 Example of RobPicker label correction on CZII tomogram TS_6_4.Original CZII annotations are shown as gray transparent spheres and corrected labels from RobPicker are shown as colored particles. Some gray transparent spheres appear smaller because they are located on tomogram slices different from the slice shown in the figure. Green arrows mark examples of recovered particle coordinates, and red arrows mark examples of removed or corrected false-positive annotations.Extended Data Fig. 2 RobPicker achieved substantially higher particle-picking F4 scores than PyTom on the CZII test dataset.Box plots show per-tomogram F4 scores for beta-galactosidase and thyroglobulin on the CZII public test subset used for this comparison (n = 50 tomograms per particle class and method). The center line denotes the median, box bounds denote the first and third quartiles, whiskers extend to 1.5 times the interquartile range, and individual points denote tomograms.Source dataSupplementary informationReporting Summary (download PDF )Peer Review File (download PDF )Source dataSource Data Fig. 1 (download XLSX )Statistical source data for Fig. 1c: DR by particle-size category.Source Data Fig. 2 (download XLSX )Statistical source data for Fig. 2a–i: D1–D5 F1 scores, summary statistics, P values, relative improvements and repetition notes.Source Data Fig. 3 (download XLSX )Statistical source data for Fig. 3: CZII public and private benchmark F4 scores.Source Data Fig. 6 (download XLSX )Statistical source data for Fig. 6: SHREC 2019 F1 scores.Source Data Extended Data Fig. 2 (download XLSX )Statistical source data for Extended Data Fig. 2: per-tomogram F4 scores for beta-galactosidase and thyroglobulin comparing RobPicker and PyTom.Rights and permissionsSpringer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.Reprints and permissionsAbout this article