Extended Data Fig. 3: Additional CellWhisperer analyses of human organ development.

a) Correlation between CellWhisperer scores (y-axis) and the expression of literature-reported organ marker genes (x-axis) for each organ across time points (dots). For both metrics, the mean across all cells at each time point was calculated and then standardized across all time points (as z-scores). Pearson correlation coefficients are reported together with their associated p-values (two-sided t-test for correlation significance). b) Number of papers per gene that co-mentioned the gene name and the organ name (as in Fig. 3c; results for heart are shown in both panels). Shared marker genes (light brown) indicate the overlap between CellWhisperer-identified (brown) and literature-reported (dark grey) organ marker genes, quantified by odds ratio and p-value (two-sided Fisher’s exact test). A size-matched set of random genes is plotted for comparison (light grey). Violin plots are shown with inner boxplots corresponding to the interquartile range and whiskers extending to the farthest data point within 1.5 times the interquartile range. c) Number of papers per gene that co-mention the gene name and the organ name (as in Fig. 3d; results for heart are shown in both panels), stratified by gene expression enrichment in CellWhisperer-identified organ-specific cells (x-axis). P-values correspond to two-sided Mann-Whitney U tests comparing CellWhisperer-identified marker genes (rightmost violin with a log₂ fold-change above 1.5) and non-marker genes (leftmost violin with log₂ fold-change below zero). Genes with strongly enriched expression in CellWhisperer-identified organ-specific cells but no associated papers are marked with red boxes. d) Gene set enrichment analysis for CellWhisperer-identified heart marker genes. The combined score (x-axis; c = log(p) × z) integrates the Fisher’s exact test p-value (two-sided) with the z-score of rank deviation to capture both statistical significance and effect size. e) Spatial gene expression in a Carnegie stage 8 (CS8) human embryo for two CellWhisperer-identified marker genes of the developing liver (bottom) and for two established liver marker genes (top). The notochord as a reference region is marked in orange, and gene expression is denoted by red points.