Fig. 5: Analyzing HCC data.

a Top, annotated regions in 4 samples (HCC-1L, HCC-2L, HCC-3L, HCC-4L): Normal, Transition, and Tumor regions. Middle and bottom, the (scaled) proportions of Fibroblasts and Endothelial cells inferred by SONAR. b Boxplots of regional proportions of Fibroblasts (top) and Endothelial cells (bottom) inferred by SONAR in 4 samples (n = 1064; 1063; 1684; 1381; 409; 1922; 1083; 601; 1281; 1685; 1311; 2127 spots from left to right boxes, respectively). c Correlations of Fibroblasts and B Cells proportions which inferred by each algorithms in the three HBV-positive samples (HCC-1L, HCC-2L, HCC-4L): Global, Normal, Transition, and Tumor regions. Color indicates two-sided Spearman rank correlations with p-values represent the significance levels: ****p ≤ 0.0001, ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05. d The spatial plots (major sub-panel) show the spatial distribution of LoCo scores of Fibroblasts and B Cells inferred by SONAR on each spots in samples HCC-1L, HCC-2L, and HCC-4L. The shapes of spots indicate the Transition region (rhombus shape) and non-Transition regions (circle shape). The line plots in the upper/left sub-panels show the marginal distribution of the (vertical/horizontal) average proportions of Fibroblasts (purple) and B Cells (blue). Vert.Mar.Prop. vertical marginal proportion, Hori.Mar.Prop. horizontal marginal proportion. e The stratified distribution of LoCo scores of Fibroblasts and B Cells based proportions inferred by SONAR (n = 781; 314; 182; 120; 78; 80; 1199; 330; 727; 790; 865; 275; 295; 1386; 877; 570; 236; 175; 108; 215; 1926 spots from left to right boxes, respectively). Spots are stratified according to the regional composition of their neighboring set into (right sub-panel): Pure Normal, Outer Edge to Transition region (Normal or Tumor side), Inner Edge to Transition region (Normal or Tumor side), Interior Transition region, and Pure Tumor region. Each box plot in b and e ranges from the first and third quartiles with the median as the horizontal line, while whiskers represent 1.5 times the interquartile range from the lower and upper bounds of the box. Source data are provided as a Source Data file including the exact p-values in c.