O-SNAP uncovers nanoscale chromatin remodeling in dedifferentiation and stress responses.

The multi-scale organization of chromatin underlies gene regulation and cell identity, yet how nuclear architecture remodels during cell state transitions remains poorly understood. Here, we use single-molecule localization microscopy and a comprehensive analytical framework we call O-SNAP to reveal distinct chromatin remodeling trajectories in two biological contexts: dedifferentiation in chondrocytes and nuclear oxidative stress-induced remodeling in mammary epithelial cells. Conventional analyses of single-molecule localization microscopy chromatin images based on qualitative inspection or simple metrics, such as chromatin domain size, fail to capture the subtle chromatin transitions in both contexts. In contrast, O-SNAP quantitatively integrates and compares 144 spatial features extracted from single-molecule localization microscopy data, allowing machine-learning based classification of nuclear states and systematic downstream analyses such as feature selection, volcano plots, and feature set enrichment analysis to determine which spatial features most strongly drive classification results. Our analysis shows that in chondrocytes, in vitro passaging drives heterochromatin formation at late passages, whereas intermediate passages exhibit heterogeneous chromatin remodeling. In contrast, in mammary epithelial cells, nuclear oxidative stress leads to chromatin decompaction specifically in cells overexpressing the oxidation-sensitive histone H3.1 variant. Together, these findings demonstrate that integrated, multiscale spatial features of chromatin are sufficient to robustly discriminate distinct cellular states.