Fig. 4: SMG progenitor cells display inflammatory and EMT programs with impaired ciliogenesis.

a Experimental schema. b Principal-component analysis (PCA) of the top 500 most variable genes expressed in the two groups. Numbers in parentheses indicate the percent variance captured by each PC. c Relative expression of Sox9 and canonical BC genes. d Heatmap of 1443 differentially expressed genes (DEGs) following unsupervised hierarchical clustering, defined by |log2FoldChange | > 0.58 and adjusted p-values < 0.05 with multiple testing correction using the Benjamini–Hochberg method (e) GSEA enrichment plots for two of the most overrepresented gene sets. Gene set enrichment analysis (GSEA) was performed using a weighted Kolmogorov–Smirnov–like statistic to calculate enrichment scores. Statistical significance was assessed by permutation testing, and nominal P values were derived from the permutation-based null distribution. Normalized enrichment scores (NES) were calculated to account for differences in gene set size, and multiple hypothesis testing was controlled using the Benjamini–Hochberg false discovery rate (FDR) correction. f–h Heatmaps of DEGs related to alarmins and inflammatory genes (f), collagen component genes (g), and ciliogenesis and cilia assembly genes (h). i, j Representative images (i) and quantification (j) of acetylated tubulin (Ac-tub) staining on tracheal sections from the indicated strains at day 14. Scale bars: 20 μm. The dashed line marks the border between SAE and SM. Group sizes: n = 7, 8 mice, from left to right. Two-tailed unpaired Mann-Whitney U test. Box plots (j) show median (center line), 25th–75th percentiles (box), and min–max values (whiskers). Each data point represents the average of 3–4 images from one mouse. See Methods. Panel a was created in BioRender. Lee, M. (https://BioRender.com/dg3i7rm). Source data are provided as a Source Data file.