Figure 1: Schematic and Calibration of EDICTS (Epi-mark Descriptor Imaging of Cell Transitional States).

(A) Workflow schematic of texture based high content image analysis of nuclear structure signatures. Undifferentiated stem cells are exposed to defined environmental cues and high resolution images of their immunolabeled nuclei are subsequently acquired via STED confocal nanoscopy. Haralick texture descriptors are extracted from these images via quantifying neighboring pixel relationships from grey-level co-occurrence matrices (GLCMs). The resolution of single pixel units correspond to the occupation of 1–4 nucleosomes. (B) Size scale relationship between high-content confocal images obtained with STED nanoscopy and electron micrographs obtained with TEM. Quantifications obtained from these images correspond to the dynamics of the gene transcription regulating bivalent mark H3K4K27me3. (C) Mass spectrometric quantification of H3K4me3 and H3K27me3 reveals their relative abundances are not significantly altered in response to soluble growth factors and differentiation. Lack of statistical significant differences determined by one-way ANOVA (K4me3: F(4,10) = 0.537235, p = 0.712008; K27me3: F(4,10) = 1.188573, p = 0.373457). (D) Correlation between LCSM and TEM quantitative datasets. A positive correlation between the “open chromatin index” from TEM micrographs and “phenotype parsing index” from LCSM images is apparent, with negative values of both indices corresponding to relatively closed chromatin states in post-mitotic cells and positive values of both indices corresponding to relatively open chromatin in undifferentiated stem cells. DZNep and MTA treated cells correspond to their positive and negative extremes, respectively. (E) Heat map of p-values from T-Tests computed between texture descriptor datasets of untreated and KMTi treated hMSCs reveals the sensitivity of all 104 texture descriptors to increasing KMTi concentration.