Extended Data Fig. 9: scChromHMM defined feature chromatin states related to ICM and TE specification.
From: Genome-coverage single-cell histone modifications for embryo lineage tracing
(a) Top, UMAP projection of gene expression of blastocysts (n = 246). Cells are defined as ICM or TE cells based on expression of ICM or TE marker genes. Bottom, expression of representative marker genes of ICM (Klf4, Etv5, Utf1, and Pou5f1) or TE (Cdx2, Id2, Gata3, and Krt18) overlaied on single cell RNA UMAP. (b) Examples of track views on representative gene loci with ChromHMM defined chromatin stages. Top: Aggregate profiles of six histone modifications in ICM and TE cells. Bottom: ChromHMM annotations. (c) Spearman correlation bewteen aggregated ICM and TE H3K4me3, H3K37me3, and H3K9me3 TACIT profiles and corresponding low-input ChIP-seq data. The correlation is calculated based on signals in genome-wide 50-kb bins. (d) Emission probabilities for histone modifications in 12 ChromHMM chromatin states, with labels on the right. (e) Receiver operating characteristic (ROC) of the random forest model. Bam files without one of the six histone modifications are used to construct the random forest model. (f) Genomic distribution of the 780 classifier bins relative to TSS. (g) Enriched GO terms of the Mouse Phenotype Single KO items for the 780 classifier bins. P value, one-sided binomial test. (h) The average Cramer’V similarity between each interpolated cell and ICM or TE cells. Interpolated cells of higher similarity to ICM or TE cells are categorized as ICM-potential or TE-potential cells, respectively. Zygote, 2cell_1, 2cell_2, 4cell_2, 8cell_1, and 8cell_2 exhibit comparable similarity to both cell types, indicating an undefined cellular potential. (i) Bar plots showing GO terms enriched in 59 ICM-related TFs (top) and 42 TE-related TFs (bottom). P value, one-sided binomial test.
