STRUCTURED CORRELATION DETECTION WITH APPLICATION TO COLOCALIZATION ANALYSIS IN DUAL-CHANNEL FLUORESCENCE MICROSCOPIC...

Area of Science:

• Microscopy and Imaging Science
• Statistical Signal Processing
• Biophysics

Background:

• Traditional colocalization analysis in fluorescence microscopy relies on user-defined regions of interest (ROIs), introducing significant analytical bias.
• Existing methods for detecting correlated regions between random processes often suffer from bias and reduced statistical power.

Purpose of the Study:

• To develop an automatic, unbiased method for structured detection of correlated regions in imaging data.
• To address the limitations of manual ROI selection and improve the reliability of colocalization analysis.
• To enhance the statistical power and accuracy of correlated region detection algorithms.

Main Methods:

• Introduced an automatic, unbiased structured detection method for correlated region detection between two random processes.
• Proposed a simple size-based normalization technique to mitigate bias and improve statistical power.
• Evaluated the method's performance in detecting structured correlations across various problems.

Main Results:

• The developed method provides automatic and unbiased detection of correlated regions.
• Size-based normalization effectively overcomes the bias and power reduction issues associated with direct maximum log-likelihood statistics.
• The proposed statistic demonstrates optimal correlated region detection capabilities for a range of structured correlation problems.

Conclusions:

• The novel automatic detection method offers a significant advancement for unbiased colocalization analysis in fluorescence microscopy.
• The size-based normalization is a key innovation for improving the robustness and accuracy of correlated region detection.
• This approach enhances the reliability of identifying spatial relationships in complex imaging datasets.