DeepPicker: A deep learning approach for fully automated particle picking in cryo-EM

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Particle picking is a critical yet time-consuming bottleneck in automated electron cryo-microscopy (cryo-EM) single-particle analysis.
• Current methods often require significant user intervention, hindering the efficiency of cryo-EM pipelines.

Purpose of the Study:

• To develop an automated particle picking framework for cryo-EM single-particle analysis.
• To reduce the manual effort and time required for particle identification in cryo-EM data processing.

Main Methods:

• Implementation of a deep learning framework named DeepPicker.
• Utilizing a novel cross-molecule training strategy to identify common particle features across different datasets.
• Testing the framework on recently published cryo-EM data from three complexes.

Main Results:

• DeepPicker successfully performs particle picking at a human-expert level.
• The framework identifies a sufficient number of particles comparable to manual picking.
• No human intervention was required during the particle picking process.

Conclusions:

• DeepPicker offers a practically useful tool to automate particle picking in cryo-EM.
• The framework significantly reduces time and manual effort in single-particle analysis.
• DeepPicker facilitates high-resolution cryo-EM structure determination and is available as open-source software.