CryoSift: an accessible and automated CNN-driven tool for cryo-EM 2D class selection.

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Single-particle cryo-electron microscopy (cryo-EM) is crucial for determining molecular structures.
• Automating data processing in cryo-EM remains a significant challenge, hindering high-throughput analysis.
• Current methods often require extensive manual intervention, limiting accessibility and speed.

Purpose of the Study:

• To develop an automated tool for assessing the quality of 2D averages in cryo-EM data.
• To enable automatic selection of suitable particles for high-resolution cryo-EM reconstructions.
• To integrate this tool into a fully automated processing pipeline for enhanced efficiency.

Main Methods:

• Development of a platform-independent convolutional neural network (CNN) named CryoSift.
• Integration of CryoSift into an automated processing pipeline utilizing cryosparc-tools.
• Implementation of a customizable 2D assessment workflow for particle selection.

Main Results:

• CryoSift effectively assesses the quality of 2D averages, facilitating automatic particle selection.
• The integrated pipeline streamlines cryo-EM data processing, reducing manual effort.
• The workflow supports high-throughput processing, benefiting both novice and experienced users.

Conclusions:

• CryoSift offers a robust solution for automating particle selection in cryo-EM.
• The developed workflow significantly improves the efficiency and accessibility of cryo-EM data analysis.
• This advancement supports the broader application of cryo-EM for high-resolution structure determination.