Fig. 2. Linking subclonal SVs to their functional consequences in LCLs.
a, Complex SVs in NA20509, with BFB-mediated rearrangements (17p) and a terminal dispersed duplication (5q) present with CF = 85%, shown for representative single cells. Ref, cells lacking complex SVs; InvDup, inverted duplication; terDel, terminal deletion. Reads are mapped to the W (orange) or C (green) strand. Gray, single cell IDs. b, Heatmap of 18 genes with altered gene activity amongst subclones, based on scNOVA (‘17p-BFB’, SV subclone; ‘17p-Ref’, 17p not rearranged). Asterisks denote TF targets of c-Myc and Max. c, Gene set over-representation analysis for TF target genes showing significant enrichment of c-Myc and Max targets in the 17p-BFB subclone. Padj., adjusted P value. Right, Model for c-Myc/Max target activation in NA20509 based on scNOVA, combined with previous knowledge. d, Mean RNA-seq expression Z scores of c-Myc/Max target genes across 33 LCLs. e, Fishplot showing CF changes over long-term culture from 23.3% (7 of 33 cells; p4) to 100% (30of 30 cells; p8). f, qPCR verifies clonal expansion of the BFB clone in p8 compared with p4 (P value based on FDR-corrected two-sided unpaired t-tests; n = 3). HG1505, control cell line with a somatically stable MAP2K3 locus. Note that for both NA20509 and HG1505 the germline copy number of the MAP2K3 locus was consistently estimated to be three. Data are presented as mean values ± s.e.m. g, RNA-seq shows significant increase of MAP2K3 at p8 versus p4 (FDR-corrected two-sided Wald test, based on DESeq2; n = 5 and three biological replicates for p4 and p8, respectively). h, Mean RNA expression Z scores of c-Myc/Max target genes in NA20509 (differences between p4 and p8 were evaluated using a two-sided Wilcoxon rank sum test; n = 5 and three biological replicates for p4 and p8, respectively). Boxplot was defined by minima = 25th percentile – 1.5 × IQR, maxima = 75th percentile + 1.5 × IQR, center = median and bounds of box = 25th and 75th percentile (g–h).