English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Cryo-Electron Tomography (cryo-ET) is a 3D imaging technology that enablesthe visualization of subcellular structures in situ at near-atomic resolution.Cellular cryo-ET images help in resolving the structures of macromolecules anddetermining their spatial relationship in a single cell, which has broadsignificance in cell and structural biology. Subtomogram classification andrecognition constitute a primary step in the systematic recovery of thesemacromolecular structures. Supervised deep learning methods have been proven tobe highly accurate and efficient for subtomogram classification, but sufferfrom limited applicability due to scarcity of annotated data. While generatingsimulated data for training supervised models is a potential solution, asizeable difference in the image intensity distribution in generated data ascompared to real experimental data will cause the trained models to performpoorly in predicting classes on real subtomograms. In this work, we presentCryo-Shift, a fully unsupervised domain adaptation and randomization frameworkfor deep learning-based cross-domain subtomogram classification. We useunsupervised multi-adversarial domain adaption to reduce the domain shiftbetween features of simulated and experimental data. We develop anetwork-driven domain randomization procedure with `warp' modules to alter thesimulated data and help the classifier generalize better on experimental data.We do not use any labeled experimental data to train our model, whereas some ofthe existing alternative approaches require labeled experimental samples forcross-domain classification. Nevertheless, Cryo-Shift outperforms the existingalternative approaches in cross-domain subtomogram classification in extensiveevaluation studies demonstrated herein using both simulated and experimentaldata.

Cryo-shift: Reducing domain shift in cryo-electron subtomograms with unsupervised domain adaptation and randomization