Fig. 2: Seq-Scope-X reveals subcellular heterogeneity in liver spatial transcriptome.
From: Seq-Scope-eXpanded: spatial omics beyond optical resolution
a–h Segmentation-based nuclear-cytoplasmic analysis. Number of segments analyzed: n = 18,448. Boxed regions are progressively magnified. a Gaussian-filtered spliced (red), unspliced (green), and mitochondrial (blue) transcripts segmented into nuclear and cytoplasmic regions using a Watershed algorithm. b Nuclear (gray) and cytoplasmic (white) segments. c Segments colored by hepatocyte zonation type: centrilobular (CL), pericentral (PC), central-side midzone (CM), portal-side midzone (PM), and periportal (PP). d–f UMAP of nuclear and cytoplasmic segments colored by cell type (d) compartment (e) or transcript count (f). g Nuclear–cytoplasmic pairs from single cells connected in UMAP space. h Heatmap showing mismatches between nuclear and cytoplasmic cell types across neighboring zonation clusters. i Segmentation-free FICTURE analysis of Seq-Scope-X liver data showing pixel-level cell type factors overlaid with nuclei. Boxed areas are magnified on the right. Arrows indicate nuclear–cytoplasmic mismatches. j FICTURE analysis of MERSCOPE liver data31 showing similar nuclear–cytoplasmic differences. Boxed areas are magnified on the right. Arrows indicate nuclear–cytoplasmic mismatches. k, l smFISH images of zonation markers with F-actin–defined cell boundaries, illustrating nuclear or cytoplasmic localization. Boxed areas are magnified on the bottom. Reproduced with permission from © 2017 Springer Nature32. All rights reserved. Scale bars represent the scale corresponding to the original tissue.
