Grotjahn Danielle A
Department of Integrative Structural and Computation Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA.
Curr Opin Struct Biol. 2025 Aug;93:103114. doi: 10.1016/j.sbi.2025.103114. Epub 2025 Jul 10.
Cryo-electron tomography provides an unprecedented view of cellular architecture, yet extracting meaningful biological insights remains challenging. Segmentation is a crucial step in this process through its ability to identify structural relationships between subcellular components visible in cryo-electron tomography data. While segmentation pipelines were historically low throughput, recent advancements in deep learning have significantly improved their automation, accuracy, and scalability. This review explores how these innovations redefine best practices for segmentation and accelerate biological discovery. This article highlights the critical role of segmentation in unlocking the full potential of cryo-electron tomography-not only for resolving macromolecular structures but also for quantifying their impact on subcellular organization and function.
冷冻电子断层扫描提供了前所未有的细胞结构视图，但提取有意义的生物学见解仍然具有挑战性。分割是这个过程中的关键步骤，因为它能够识别冷冻电子断层扫描数据中可见的亚细胞成分之间的结构关系。虽然分割流程在历史上通量较低，但深度学习的最新进展显著提高了它们的自动化程度、准确性和可扩展性。本综述探讨了这些创新如何重新定义分割的最佳实践并加速生物学发现。本文强调了分割在释放冷冻电子断层扫描的全部潜力方面的关键作用——不仅用于解析大分子结构，还用于量化它们对亚细胞组织和功能的影响。
