Leopard-EM: an extensible 2D template-matching package to accelerate in situ structural biology

Area of Science:

• Structural biology
• Cellular imaging
• Computational microscopy

Background:

• Cryogenic electron microscopy (cryo-EM) provides high-resolution cellular views, revealing molecular mechanisms.
• Accurate annotation of structures in cryo-EM images is challenging, limiting the technique's impact.
• 2D template matching (2DTM) shows promise for locating and orienting molecular complexes in cryo-EM data.

Purpose of the Study:

• To develop a modular and extensible software architecture for 2D template matching (2DTM).
• To accelerate innovation in cryo-EM data processing and analysis.
• To improve the accuracy and efficiency of annotating molecular structures in cellular cryo-EM images.

Main Methods:

• Developed Leopard-EM, a modular Python-based implementation of 2DTM.
• Implemented an automated pixel-size refinement procedure, identifying sensitivity to errors of ~0.2%.
• Created a constrained search protocol utilizing initial particle locations for enhanced detection.

Main Results:

• Leopard-EM offers a customizable and extensible platform for 2DTM.
• The automated pixel-size refinement is crucial for accurate 2DTM.
• A constrained search protocol improved small ribosomal subunit detection eightfold.
• Captured ribosome rotation states within living cells at single-molecule resolution.

Conclusions:

• Leopard-EM facilitates the development of novel 2DTM approaches for in situ cryo-EM.
• The software supports rapid innovation in cryo-EM data processing workflows.
• Leopard-EM enables high-resolution analysis of molecular mechanisms within their native cellular environment.