Density-Aware Spatial Randomization (DenSR) generates realistic null models for object-based colocalization analysis.

Area of Science:

• Cell Biology
• Biophysics
• Microscopy

Background:

• Quantitative colocalization analysis in fluorescence microscopy is crucial for studying molecular interactions.
• Object-based methods face challenges distinguishing true associations from random overlap in heterogeneous cellular environments.
• Existing randomized null models often fail to preserve spatial density, leading to inaccurate colocalization estimates.

Purpose of the Study:

• To introduce Density-aware Spatial Randomization (DenSR), a novel framework for generating spatially realistic null models.
• To improve the accuracy of colocalization estimation by accounting for heterogeneous spatial distributions of biomolecules.
• To enhance the sensitivity of detecting true molecular interactions.

Main Methods:

• Development of the DenSR in silico framework.
• Generation of density-aware null models that preserve local clustering and global spatial organization.
• Application of DenSR to protein and RNA datasets with varying spatial heterogeneity.

Main Results:

• Uniform randomization significantly underestimates background proximity in heterogeneous environments.
• Density-aware null models provide more accurate estimations of random colocalization.
• DenSR prevents overestimation of colocalization caused by underestimated background proximity.
• Different randomization strategies converge for sparse transcript datasets.

Conclusions:

• DenSR offers a robust approach for estimating chance colocalization across diverse spatial distributions.
• The framework enhances the reliability of quantitative colocalization analysis in fluorescence microscopy.
• Accurate null models are essential for sensitive detection of molecular interactions in complex biological systems.