Subcellular structure segmentation from cryo-electron tomograms via machine learning

Abstract

We describe how to use several machine learning techniques organized in a learning pipeline to segment and identify subcellular structures from cryo electron tomograms. These tomograms are difficult to analyze with traditional segmentation tools. The learning pipeline in our approach starts from supervised learning via a special convolutional neural network trained with simulated data. It continues with semi-supervised reinforcement learning and/or a region merging techniques that try to piece together disconnected components that should belong to the same subcellular structure. A parametric or non-parametric fitting procedure is then used to enhance the segmentation results and quantify uncertainties in the fitting. Domain knowledge is used in generating the training data for the neural network and in guiding the fitting procedure through the use of appropriately chosen priors and constraints. We demonstrate that the approach proposed here work well for extracting membrane surfaces of protein reconstituted liposomes in a cellular environment that contains other artifacts.April

Quantifying organellar ultrastructure in cryo-electron tomography using a surface morphometrics pipeline

“…The majority of these data were collected without the use of a post-column energy filter ( Koning et al, 2018 ) or phase plate ( Imhof et al, 2019 ), both of which improve contrast of visible features within tomograms. Additionally, there have been recent advances in algorithmic approaches to improve the quality and completeness of segmentations ( Zhou et al, 2020 Preprint ; Chen et al, 2017 ; Buchholz et al, 2018 ; Liu et al, 2021 Preprint ). The flexibility of the screened Poisson surface reconstruction workflow will enable easy adaptation of these morphometric approaches to new segmentation tools.…”

A surface morphometrics toolkit to quantify organellar membrane ultrastructure using cryo-electron tomography

“…The majority of these data were collected without the use of a post-column energy filter 45 or phase plate 46 , both of which improve contrast of visible features within tomograms. Additionally, there have been recent advances in algorithmic approaches to improve the quality and completeness of segmentations 14,[47][48][49] . The flexibility of the screened Poisson surface reconstruction workflow will enable easy adaptation of these morphometric approaches to new segmentation tools.…”

Quantitative Cryo-Electron Tomography

“…One-shot learning 3D segmentation explores this possibility by performing 3D segmentation in simulated cryo-ET data based on deep learning approaches trained on a single sample set per class. Interestingly machine learning workflows aiming to combine the improvement of image contrast, 2D segmentation, refinement using reinforced learning, classification, and 3D refinement of the surface ( Zhou et al, 2020 ) appear as promising tools for quantitative cryo-ET.…”