Extended Data Fig. 4: SIMPLE-seq generates cell-type-specific 5mC and 5hmC profiles from human PBMC.

a, Heatmap showing the gene expression and promoter methylation levels of marker genes for the five major cell types. b and c, UMAP embedding showing the single-cell clustering based on (b) 5mCHG and (c) 5hmCHG levels (in 100-kb non-overlapping bins) from PBMC. Each dot represents a single cell and is colored according to its annotation based on 5mCG.d, Silhouette plot showing the degree of separation of the PBMC clusters based on 5mC, 5hmC or joint 5mCG-5hmCG. e, UMAP embedding showing the single-cell clustering based on joint 5mCG-5hmCG levels (in 100-kb non-overlapping bins) from PBMC. Each dot represents a single cell and is colored according to its annotation based on 5mCG. f and g, UMAP embedding showing the single-cell clustering based on (f) 5mCG levels of H3K4me1 regions (g) 5hmCG levels of H3K4me1 regions from PBMC. Each dot represents a single cell and is colored according to its annotation based on 5mCG. h, Violin plots showing the modification levels of 5mCG, 5hmCG, 5mCHG and 5hmCHG in different cell types. i, The heatmap showing the numbers of pairwise differentially methylated (5mCG) regions across cell types (in 5-kb non-overlapping bins). j, Heatmap showing the methylation levels of 5mCG DMRs of a representative group (CD4⁺ T cells and Monocytes). k-m, The enrichment analysis of (k) know motifs, (l) top enriched de novo motifs, P-value, one-sided Fisher’s exact test, and (m) top enriched GREAT GO terms, P-value, one-sided Fisher’s exact test.