Measuring the similarity of single-molecule localization microscopy derived marked point clouds

Cellular membranes are dynamic, heterogeneous structures where lipid nanodomains (e.g., lipid rafts) play key roles in signaling, membrane trafficking, and protein function. Single-molecule localization microscopy (SMLM) reveals the spatial organization of these nanoscale features; however, traditional analyses focus only on spatial patterns and neglect biochemical and biophysical properties critical for membrane function. By combining SMLM with environmentally sensitive fluorescent probes, such as di-4-ANEPPDHQ, we can produce marked point patterns that couple spatial coordinates with environmental information, such as membrane lipid order quantified by generalized polarization (GP) values. Unfortunately, existing methods do not adequately compare these complex data sets. Here, we introduce a new method, which assesses the similarities of marked point patterns by considering the spatial arrangement as well as the biophysical properties of the data. The method computes three semi-independent Kolmogorov-Smirnov scores which are used to map comparisons between two point clouds in 3D space. This allows the distance to the origin of a comparison to be used as a metric for similarity. Application to simulated data confirms the reliability of the method, while application to experimental GP-marked point patterns identifies condition-dependent variations in lipid order. This framework thus offers a versatile tool for the study of biochemical and biophysical properties of cellular nanoenvironments, enabling new insight into membrane organization and function.