A convolutional autoencoder approach for mining features in cellular electron cryo-tomograms and weakly supervised coarse segmentation

• Published in: Journal of structural biology Vol. 202; no. 2; pp. 150 - 160
• Main Authors: Zeng, Xiangrui, Leung, Miguel Ricardo, Zeev-Ben-Mordehai, Tzviya, Xu, Min
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01-05-2018
• Subjects: Algorithms; Cellular electron cryo-tomography; Convolutional autoencoder; Convolutional neural network; Cryoelectron Microscopy - methods; Deep learning; Image Processing, Computer-Assisted - methods; Image semantic segmentation; Machine learning; Macromolecular complex; Macromolecular Substances - chemistry; Particle picking; Pose normalization; Software; Structural pattern mining; Subtomogram classification; Unsupervised learning; Visual proteomics; Weakly supervised learning; Image semantic segmentation; Visual proteomics; Cellular electron cryo-tomography; Subtomogram classification; Convolutional neural network; Convolutional autoencoder; Macromolecular complex; Unsupervised learning; Deep learning; Particle picking; Weakly supervised learning; Machine learning; Pose normalization; Structural pattern mining
• ISSN: 1047-8477; 1095-8657
• DOI: 10.1016/j.jsb.2017.12.015

Cellular electron cryo-tomography enables the 3D visualization of cellular organization in the near-native state and at submolecular resolution. However, the contents of cellular tomograms are often complex, making it difficult to automatically isolate different in situ cellular components. In this paper, we propose a convolutional autoencoder-based unsupervised approach to provide a coarse grouping of 3D small subvolumes extracted from tomograms. We demonstrate that the autoencoder can be used for efficient and coarse characterization of features of macromolecular complexes and surfaces, such as membranes. In addition, the autoencoder can be used to detect non-cellular features related to sample preparation and data collection, such as carbon edges from the grid and tomogram boundaries. The autoencoder is also able to detect patterns that may indicate spatial interactions between cellular components. Furthermore, we demonstrate that our autoencoder can be used for weakly supervised semantic segmentation of cellular components, requiring a very small amount of manual annotation.