Extended Data Fig. 5: PerturbView in tissue.
From: Multiplexed, image-based pooled screens in primary cells and tissues with PerturbView
a–d, High barcoding decoding accuracy in tissue sections with PerturbView. a, Barcode mapping rate (y axis, fraction of assigned reads at a Hamming distance ≤1 to the pre-defined sgRNA lookup table) for conventional ISS and PerturbView (x axis) in FF and FFPE tumor tissue (n = 4). b, Mean percentage of cells with detectable sgRNA (y axis) captured by conventional ISS or by PerturbView (x axis) in FFPE (light green) or fresh frozen (FF) (dark green) subcutaneous tumors. n = 2 independent experiments, Error bars: SD. c, Representative image of immunofluorescence stain with anti-human PRKDC in DLD-1 xenograft tissue section. Mouse cells (outer rim) or necrotic regions (central region) show low PRKDC staining. Scale bar, 1 mm. d, Frequency of each sgRNA (dots; normalized counts) estimated by NGS of the plasmid library (x axis) or PerturbView (y axis) of FF (dark green) or FFPE (light green) samples. n = 2 for each FF and FFPE treatment. Pearson’s r is noted. e, Spatial transcriptome and barcode detection in UMAP and spatial domains. UMAP embedding on transcriptome profiles (dots) colored and numbered by Leiden clusters (far left); and tissue section colored by the same Leiden clusters (second from left), total transcript counts in each cell (second from right) and minimum barcode hamming distance (up to 2, right); and distribution of total transcript counts per cell (x axis) for each Leiden cluster (y axis). Interior box plots indicate the median (white lines), interquartile range (hinges) and 1.5 times of the interquartile range (whiskers). (n = 2 consecutive sections per tumor). f, g, Expression changes associated with clonal diversity. f, Tumor sections profiled (scale bar, 1 mm), colored by Shannon diversity (top) or corresponding diversity groups (bottom) for all analyzed sections aside from the one shown in Fig. 3g. g, Genes (columns) that are differentially expressed (log fold-change > 0.5) between regions with high and low Shannon diversity (rows) for each tumor section (n = 2 consecutive sections per tumor).
