Extended Data Fig. 1: Quality control and benchmarking of CHARM.
From: Gene regulatory landscape dissected by single-cell four-omics sequencing
a,b, Scatter plots showing quality control metrics used for filtering CHARM data. a, Fraction of reads in peaks (FRiP) for H3K27me3 versus chromatin accessibility. b, Number of Hi-C contacts versus RNA UMIs per cell. Dashed lines indicate filtering thresholds. c, Violin plots showing TSS enrichment per cell across different methods. d-g, Violin plots comparing per-cell data yield from CHARM with other single-cell methods across four modalities: d, RNA UMIs; e, H3K27me3 fragments; f, accessible chromatin fragments; and g, Hi-C contacts. Boxes indicate the median and 25th-75th percentiles, with whiskers extending to 1.5× the interquartile range in panel c-g. h, Scatter plot of Sox2 UMI counts versus enhancer–promoter (E–P) interaction strength, with cells ordered by Sox2 expression. Each dot represents one cell (n = 720 cells). Pearson’s correlation coefficient is indicated. i, Box plots comparing E–P interaction strength between cells with high and low Sox2 expression. Cells were divided into high (n = 176 cells) and low (n = 544 cells) Sox2 expression groups. Boxes indicate the median and 25th-75th percentiles, with whiskers extending to 1.5× the interquartile range. Statistical significance was assessed using a two-sided Wilcoxon rank-sum test (P = 0.72).
