DeepSTORM3D: dense 3D localization microscopy and PSF design by deep learning

• Published in: Nature methods Vol. 17; no. 7; pp. 734 - 740
• Main Authors: Nehme, Elias, Freedman, Daniel, Gordon, Racheli, Ferdman, Boris, Weiss, Lucien E., Alalouf, Onit, Naor, Tal, Orange, Reut, Michaeli, Tomer, Shechtman, Yoav
• Format: Journal Article
• Language: English
• Published: New York Nature Publishing Group US 01.07.2020; Nature Publishing Group
• Subjects: 631/1647/245/2225; 631/1647/328/2238; Bioinformatics; Biological activity; Biological Microscopy; Biological Phenomena; Biological Techniques; Biomedical and Life Sciences; Biomedical Engineering/Biotechnology; Cable television broadcasting industry; Deep Learning; Design; Emitters; Evaluation; Imaging, Three-Dimensional - methods; Life Sciences; Localization; Machine learning; Microscope and microscopy; Microscopy; Mitochondria; Neural networks; Neural Networks, Computer; Optical components; Point spread functions; Proteomics; Single Molecule Imaging - methods; Telomere - ultrastructure; Telomeres; Temporal resolution; United States; Israel
• ISSN: 1548-7091; 1548-7105
• DOI: 10.1038/s41592-020-0853-5

An outstanding challenge in single-molecule localization microscopy is the accurate and precise localization of individual point emitters in three dimensions in densely labeled samples. One established approach for three-dimensional single-molecule localization is point-spread-function (PSF) engineering, in which the PSF is engineered to vary distinctively with emitter depth using additional optical elements. However, images of dense emitters, which are desirable for improving temporal resolution, pose a challenge for algorithmic localization of engineered PSFs, due to lateral overlap of the emitter PSFs. Here we train a neural network to localize multiple emitters with densely overlapping Tetrapod PSFs over a large axial range. We then use the network to design the optimal PSF for the multi-emitter case. We demonstrate our approach experimentally with super-resolution reconstructions of mitochondria and volumetric imaging of fluorescently labeled telomeres in cells. Our approach, DeepSTORM3D, enables the study of biological processes in whole cells at timescales that are rarely explored in localization microscopy. DeepSTORM3D uses deep learning for accurate localization of point emitters in densely labeled samples in three dimensions for volumetric localization microscopy with high temporal resolution, as well as for optimal point-spread function design.