Fig. 2

Inter-individual transcriptomic variability analysis (A) Sequencing of sibling embryos enables the study of inter-individual gene expression variance. A UMAP of the single-cell data without batch correction (12 hpf, four embryos, 6,699 cells total). Example of normalized expression of gene ybx1 for four embryos (color coded). (B) We used the Kullback-Leibler divergence (KLD) to quantify the pairwise similarity between probability distribution illustrating variant (top, KLD ≫ 0) and invariant (bottom, KLD ≈ 0) gene expression across embryos. (C) Example distributions for embryo-invariant genes (KLD ≈ 0) at 12 hpf: sox11a, notch1a, and dnaja2a. The x axis indicates the log-normalized expression values, and the y axis shows the density (relative number of cells). Heatmaps show all pairwise KLD values. We summarized these heatmaps per gene by the median pairwise KLD (mpKLD, values above the heatmap). (D) Example distributions for embryo-variant genes (KLD ≫ 0) at 12 hpf: ybx1, tma7, and metap2b. (E) Histograms of mpKLD values scores for all cells at 12 hpf. A few examples of high and low mpKLD genes are shown. (F) Distributions of mpKLD values across all genes at 10, 12, 14, 16, 19, and 24 hpf. In (E) and (F), the top and bottom 2% of genes are indicated as colored areas (p values < 10−15 are marked with an asterisk). See also Figure S2 and Data S1.