CryoAlign2: efficient global and local Cryo-EM map retrieval based on parallel-accelerated local spatial structural...

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Advancements in Cryo-Electron Microscopy (Cryo-EM) generate numerous 3D density maps, necessitating efficient methods for structure similarity retrieval.
• Existing tools like CryoAlign offer accurate density map alignment but lack efficiency in local alignment and retrieval capabilities.

Purpose of the Study:

• To develop an efficient and accurate alignment-based retrieval tool for Cryo-EM density maps.
• To enable both global and local structure similarity searches within large Cryo-EM datasets.

Main Methods:

• Developed a retrieval tool utilizing parallel-accelerated CryoAlign for high-precision 3D alignment.
• Transformed density maps into point clouds for efficient storage and retrieval.
• Implemented a multi-dimension scoring function to assess structural similarities.

Main Results:

• Achieved up to a 7-fold speedup in retrieval tasks compared to existing methods.
• Demonstrated high precision in local alignment and effective retrieval even when one map is contained within another.
• Successfully performed global, local, and hybrid similarity retrieval tasks.

Conclusions:

• The developed tool provides an efficient and accurate solution for Cryo-EM density map similarity search.
• Enables researchers to better utilize the growing public repository of Cryo-EM structural data.
• The tool supports precise local alignments, addressing a key limitation of previous methods.