MultiMatch: geometry-informed colocalization in multi-color super-resolution microscopy

Area of Science:

• Biophysics
• Computational Biology
• Microscopy

Background:

• Super-resolution microscopy enables nanoscale visualization of biological structures.
• Accurate colocalization analysis requires integrating prior knowledge of local complex geometry.
• Existing methods struggle with incomplete labeling and complex spatial arrangements.

Purpose of the Study:

• To introduce MultiMatch, a novel computational tool for analyzing particle arrangements in multicolor super-resolution microscopy.
• To develop an object-based colocalization model that accounts for incomplete labeling efficiencies.
• To provide a user-friendly Python package for advanced colocalization analysis.

Main Methods:

• Utilizes multi-marginal optimal unbalanced transport methodology for analyzing chain-like particle arrangements.
• Employs an object-based colocalization model to statistically address incomplete labeling.
• Applies the method to three-color STED images of DNA origami nanorulers.

Main Results:

• MultiMatch successfully recovers existing chain structures in complex microscopy data.
• The method demonstrates superior performance compared to geometry-uninformed triplet colocalization approaches.
• The tool generalizes to an arbitrary number of color channels.

Conclusions:

• MultiMatch offers a robust solution for analyzing spatial proximity in multicolor super-resolution microscopy.
• The developed methodology improves the inference of particle arrangements, even with labeling limitations.
• The Python package facilitates advanced colocalization analysis and visualization for researchers.