Fig. 3: Avian transcriptional bursting and Z chromosome upregulation.

a Scatterplot of number of biallelically expressed genes (average FPKM > 20, 1033-3579 genes) per cell (n = 1382), by cell size (FSC-A) based on FACS and single-cell RNA-seq (Smart-seq3) data. Point colours are denoted as genes with FPKM > 1 per cell. b Boxplots of allele-resolved spike-in normalised UMI counts of single-cell RNA-seq (Xpress-seq) of CEFs per sample and allele. Number of single cells (n) shown in the respective panels. Only expressed genes (average FPKM > 1, chrA = 8467, chrZ = 424 genes) were included in the analysis. Data shown as median (center line), first and third quartiles (box limits), and 1.5x interquartile range (whiskers). c Scatter plot of burst frequency (y-axis) and burst size (x-axis) inferred for expressed genes (7392 genes, average FPKM > 1) in CEFs, coloured based on FPKM-normalised expression level. d Heatmaps of core promoter elements and their correlation with burst frequency (left) or burst size (right) for expressed genes (7392 genes, average FPKM > 1). e Allele-resolved, log2-relative burst frequency per cell for autosomal (chrA = 4586) and Z-linked (chrZ = 229) genes, plotted as median ± 95% CI. RJF and WL alleles are shown in red and blue, respectively. P-values from two-tailed Mann-Whitney U tests are shown. f Relative expression levels of autosomal and X-linked genes in mouse and Z-linked genes in chicken. Histogram of medians of relative expression in FPKM of randomly subsampled genes, compared to the median of chrX for mouse (chrA genes = 10025 genes, chrX = 276 genes) or chrZ for chicken (chrA = 7180 genes, chrZ = 332 genes), based on scRNA-seq in mouse fibroblasts⁵⁰ and CEFs, with the number of cells used shown in black. P-values denote a one-tailed bootstrap test with 1×10⁵ permutations (no adjustments for multiple comparisons). C57 (C57BL6/J) and CAST (CAST/EiJ) denotes mouse alleles.