Fig. 8
Spatial association of MSGs, TE categories, and chromatin states. a Hierarchical clustering for normalized enrichment scores of features based on Spearman’s ρ in non-overlapping, 50-kb intervals. b Spatial correlation measured by relative distance between, on the one hand, CMT2- and RdDM-targeted TEs and, on the other hand, genes that are metastable (left) and nearby certain chromatin states (right). c MET and JSD profiles for rosette leaves in each context across CMT2- and RdDM-targeted TEs. d Probability of metastable genes (MSGs) nearby CMT2- and RdDM-targeted TEs. The x-axis gives the gene count normalized to the total number of MSGs (top 5%, here for rosettes). The vertical lines show the observed fraction of MSGs nearby the targeted TE category. The kernel density estimates show the distribution of the fraction of genes nearby TEs for 10,000 randomly selected gene sets, each of a size equal to the total number of MSGs. The p value that is estimated by the randomization test is defined by p=P(g≥gobserved∣H0), where the null hypothesis H0 is random selection of protein-coding genes that are not TEGs