Fig. 4: Joint analysis of 11 breast cancer ST slices from three different datasets.
From: SPACEL: deep learning-based characterization of spatial transcriptome architectures
a Spatial domains identified by Splane in slice S2 and S5 from Wu et al. dataset, slice S10 from Zhao et al. dataset, and slice S11 released by 10X Genomics. b, c Spatial distribution of chromosome 1q&8q copy number gains (b) and 1p copy number losses (c) of ST spots in slices S11, calculated by inferCNV. Dashed lines represent the tumor domain. d, e CNVs of chromosome 1q & 8q (d) and chromosome 1p (e) in each spatial domain calculated by inferCNV. CNVs, copy number variations; center line, median value; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; n = 11 slices. f From left to right: Splane predicted spatial domains in slice S5, distribution of Splane predicted immune domains D7/D8/D9, distribution of Spoint predicted immune cells, and distribution of H&E staining marked immune spots. g Percentage of H&E staining marked immune spots in each domain of slice S1, S2, S5, and S6. The four slices were H&E stained in the original study. Bar height, mean value; whiskers, mean values ± 95% confidence intervals; n = 4 slices. h From left to right: Splane predicted spatial domains in slice S10, distribution of Splane predicted immune domains D7, D8, and D9, distribution of Spoint predicted immune cells, and distribution of CD3+ immunofluorescence (IF) staining marked immune spots. i Percentage of CD3+ IF staining marked immune spots in each domain of slice S10. Source data are provided as a Source Data file.
