Fig. 3: Lineage tracing reveals SCLC trajectories.
From: Basal cell of origin resolves neuroendocrine–tuft lineage plasticity in cancer
a, Schematic of CellTagging in RPM and RPMA basal organoids and allografts. CellTagged organoids (pre-Cre and post-Cre) were collected for scRNA-seq at implant time; n = 1 RPMA, n = 1 RPM (pre-Cre) and n = 2 RPM (post-Cre) tumours (representing independent experimental replicates) were sequenced for clonal analyses. b, Fluorescence images of transformed CellTagged RPM organoids (similar results obtained in n = 4 independent CellTagging experiments). c, ForceAtlas2 (FA) map of RPM and RPMA tumours (from Fig. 2f,h) by Leiden cluster per genotype (top, Supplementary Table 3) or fate (bottom). d, Leiden cluster frequencies per clone. Each bar represents one clone. Clonal patterns, genotype and CellTag method are shown on the x axis. Individual clones are shown in Extended Data Fig. 6b,c and CellTag annotations in Supplementary Table 4. e, ForceAtlas2 maps of main RPM and RPMA clonal patterns in d. f, ForceAtlas2 maps of main clonal patterns in e, annotated by SCLC fate in c. g, ForceAtlas2 map coloured by pseudotime (start = basal-enriched cluster 17). h, ForceAtlas2 maps of main clonal patterns in e, annotated by pseudotime in g and fate in c. Straight arrows denote state transitions; circular arrows denote self-renewal within a fate. i, CellRank plots of fate probabilities in RPM and RPMA tumour cells in c, annotated by SCLC fate (left) or Leiden cluster (right). Cells arranged inside the circle according to fate probability, with fate-biased cells next to their corresponding edge and naive cells in the middle. j, CellRank-predicted expression trends of putative driver genes (Supplementary Table 5), plotted along pseudotime trajectories from basal to indicated fates/Leiden clusters. Scale bars: 650 μm (b, left), 275 μm (b, right). Schematic in a was created using BioRender (https://biorender.com).
