Fig. 2: Integrated chromatin accessibility and gene expression data at single-nucleus resolution via SUM-seq characterize differentiation trajectories along M1/M2 macrophage polarizations.

a, A schematic overview of the macrophage polarization experiment. hiPS cell-derived macrophages were stimulated with LPS and IFN-γ (M1) or IL-4 (M2). Nuclei were fixed with glyoxal and collected for SUM-seq at 0 h, 1 h, 6 h, 10 h and 24 h (n = 2 per time point). b,c, WNN UMAP projection of integrated SUM-seq data of macrophage polarization. Cells are annotated and labeled according to their sample index (arrows were drawn by hand for ease of interpretation) (b), and annotated with AUC score of M1 signature genes (c, left) and M2 signature genes (c, right; Supplementary Table 2). d,e, Distributions of cells from each time point across three MOFA factor weights associated with M1 polarization (d) (early response, sustained response and late response) and M2 polarization (e) (early response and sustained responses I and II). The dashed arrows indicate the direction of the time-resolved response. f,g, GSEA (Methods) for M1 (f) and M2 (FDR-corrected P values (Padj)) (g) polarization factors are shown as heat maps. h, Top: motif activity (rank-normalized chromVAR z scores) for TFs associated with M1 early response (Methods) across M0 and M1–1 h cells sorted by M1 early response factor weights. Bottom: motif activity for TFs associated with M1 sustained response across all M0 and M1 cells sorted by sustained factor weights. STAT1.0 and STAT1.1 represent the homodimer and ISGF3 motifs of STAT1, respectively. i, Top: motif activity for TFs associated with M2 early response across M0 and M2–1 h cells sorted by M2 early response factor weights. Bottom: motif activity for TFs associated with M2 sustained response II across all M0 and M2 cells sorted by sustained factor II weights.