Fig. 4: Single-cell transcriptional hallmarks of ageing covary with cryptic mutations.

a Volcano plot of DEGs in cells with cryptic mutations which are not synonymous above 10% heteroplasmy, and those without, in the human pancreas data³⁰. Differential expression analysis was done using the Wilcoxon rank-sum test. b Genes whose expression covaries with the presence of mutations form a functional protein-interaction module which is enriched with markers of stress response. To obtain the functional module, we combine the p values of differential expression with the curated protein interaction network from BioGRID by computing a maximum-weight connected subgraph with scPPIN⁵². c DEGs in cells with cryptic mutations which are not synonymous for human, mice, and rat. A number of cells in each class is given in Supplementary Table S6. Differential expression analysis was done using the Wilcoxon rank-sum test. d Selected GO terms that are enriched in the DEGs of at least six of the seven datasets (two human pancreas, human retina, mouse pancreas, mouse liver, rat liver, rat brown adipose tissue, see Table 1 for details). We order the GO terms by the false discovery rate (FDR) in the data from ref. ³⁰.