A Persistent Homology Approach Enables Single-Cell Detection of Chromatin Spatial Organization Changes During Pathological Senescence Transition.

Chromatin organization changes during aging, accompanied by alterations in higher-order chromosome architecture, contributing to the acquisition of senescent cell-specific functions. Recent studies have revealed that chromatin states and their regulatory mechanisms vary between individual cells. However, analyses of higher-order chromosome structure have been performed mostly at the population level. Therefore, approaches capable of directly quantifying such alterations at the single-cell level are required. Here, we applied persistent homology (PH), a topological data analysis framework, to quantify spatial features of γH2AX foci and DAPI-stained chromatin in single-cell fluorescence images. Using a model of pathological senescence induced by impaired acetylation-dependent histone H2AX exchange, we extracted topological features followed by machine learning analysis. Unsupervised analysis revealed spatial patterns associated with pathological senescence in both γH2AX and DAPI signals. Notably, classifications derived from γH2AX foci showed correspondence with those from DAPI-based chromatin organization, suggesting that local DNA damage patterns reflect global chromatin structure. Supervised models further distinguished normal and pathological senescent cells, with DAPI-derived features highlighting the contribution of nuclear intensity gradients. These findings demonstrate that PH enables quantitative characterization of nuclear architecture at the single-cell level and provides a framework for dissecting chromatin reorganization during cell state transitions.