Automatic cryo-EM particle selection for membrane proteins in spherical liposomes

Area of Science:

• Structural biology
• Biophysics
• Cryo-electron microscopy

Background:

• Single-particle cryo-EM requires accurate particle detection for structure determination.
• Membrane proteins in lipid vesicles (liposomes) pose unique challenges for particle identification.

Purpose of the Study:

• To develop an efficient algorithm for detecting membrane protein particles in cryo-EM micrographs.
• To improve the accuracy of particle selection in random spherically constrained (RSC) single particle reconstruction.

Main Methods:

• Implemented an efficient algorithm for computing position-dependent normalized cross-correlation (NCC).
• Utilized geometric constraints of particles embedded in spherical vesicle membranes.
• Applied the method to select membrane protein particles, specifically GluA2 glutamate receptors.

Main Results:

• Demonstrated an efficient method for calculating position-dependent NCC.
• Successfully applied the algorithm to identify membrane protein particles.
• Showcased the method's effectiveness with GluA2 glutamate receptors exhibiting diverse views.

Conclusions:

• The developed algorithm enhances particle detection in RSC single particle reconstruction.
• This method facilitates more reliable structure determination of membrane proteins in liposomes.
• The approach is effective even for proteins with highly variable projection directions.