Conformational heterogeneity and probability distributions from single-particle cryo-electron microscopy

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Single-particle cryo-electron microscopy (cryo-EM) captures projection images of biomolecules at cryogenic temperatures.
• A key advantage of cryo-EM is its capacity to image biomolecules in heterogeneous conformations.
• Analyzing heterogeneous cryo-EM data presents significant computational challenges.

Purpose of the Study:

• To review current methods for reconstructing and analyzing heterogeneous conformations from cryo-EM data.
• To provide an overview of dimensionality reduction techniques applied to heterogeneous 3D reconstruction.
• To discuss methods for estimating conformational probability distributions from cryo-EM images.

Main Methods:

• Review of linear-transformation-based methods for cryo-EM data analysis.
• Overview of nonlinear deep generative models for cryo-EM reconstruction.
• Examination of dimensionality reduction techniques in heterogeneous 3D reconstruction.

Main Results:

• Algorithmic advancements enable the recovery of heterogeneous conformations from noisy cryo-EM data.
• Various methods infer different types of information from cryo-EM imaging data.
• Methods are available to estimate probability distributions over conformational states.

Conclusions:

• Recent algorithmic progress facilitates the analysis of conformational heterogeneity in cryo-EM.
• A range of computational approaches, from linear to deep learning models, are applicable.
• Ongoing challenges remain for the cryo-EM community in data analysis and reconstruction.