Quantitative Network Classification with CLaSSiNet Reveals Nanoscale Organizational Principles of the Membrane...

Quantitative analysis of nanoscale biological network architectures remains challenging in super-resolution fluorescence imaging, where cytoskeletal and network-like structures often appear as sparse localization clusters rather than continuous filaments. We introduce CLaSSiNet, a first-of-its-kind modular computational pipeline that integrates connectivity, 1D periodicity, and 2D regularity classifiers to automatically segment, classify, and map nanoscale network architectures with unprecedented classification resolution (~256 nm) in single-molecule localization microscopy datasets. CLaSSiNet uniquely distinguishes four organizational states (1D periodic networks, 2D polygonal ordered networks, disordered networks, and non-network states) and is robust to experimental noise and compatible with both node- and link-based labeling strategies. Using CLaSSiNet, we achieve the first spatially resolved, quantitative mapping of organizational heterogeneity in the actin-spectrin membrane-associated periodic skeleton (MPS), a broadly conserved membrane scaffold in animal cells. We uncover previously unrecognized subcellular patterning, with ordered 1D and 2D networks enriched at cell edges and junctions, while non-network states dominate the cell body. CLaSSiNet further reveals mechanical coupling principles: actin stress fibers bias the formation and orientation of 1D periodic networks and enrich nearby non-network regions, suggesting coordination between spectrin lattices and contractile actin bundles. Comparative analysis across cell types shows that neurons favor 1D periodic architectures in neurites and 2D polygonal networks in somas, whereas U2OS and 3T3 cells adopt these architectures to a lesser extent, highlighting cell-type-specific tuning of spectrin network design. Together, these results establish CLaSSiNet as a generalizable platform for quantitative network-state mapping and provide new biological insights into how spectrin-based architectures adapt to local mechanical environments.