AlignOT: An Optimal Transport Based Algorithm for Fast 3D Alignment With Applications to Cryogenic Electron...

Area of Science:

• Structural biology
• Computational biology
• Biophysics

Background:

• Cryogenic electron microscopy (cryo-EM) is vital for determining biomolecular structures.
• Aligning cryo-EM maps is essential for analyzing conformational changes.
• Current alignment methods can be computationally expensive and may converge to suboptimal solutions.

Purpose of the Study:

• To introduce AlignOT, a novel computational method for aligning 3D cryo-EM maps.
• To address the limitations of existing local alignment techniques.
• To provide a more robust and accurate approach for 3D map alignment.

Main Methods:

• Utilizing computational optimal transport (OT) for 3D map alignment.
• Implementing a fast OT map estimation within a stochastic gradient descent algorithm.
• Minimizing Wasserstein distance between point cloud representations of EM maps.

Main Results:

• AlignOT demonstrates improved precision and accuracy in cryo-EM map alignment.
• The method shows a wider range of effective rotation angles compared to standard methods.
• Benchmarking on diverse experimental maps confirms AlignOT's robust performance across different conformational heterogeneities.

Conclusions:

• AlignOT offers a superior alternative for aligning 3D cryo-EM maps.
• The method effectively overcomes local minima issues present in traditional alignment tools.
• AlignOT is anticipated to have broad applicability in structural biology research involving 3D EM map analysis.