Fig. 2

The Xi bipartite structure is captured by allelic sci-Hi-C data and changes with the cell cycle. A Aggregate (pseudobulk) allelic contact maps for the Xi and Xa using all segregated contact read pairs from sci-Hi-C on 4493 individual Patski cells (XCI is completely skewed to the BL6 allele in Patski cells). Color scale in log10 (contact counts + 1). See also Additional file 1: Fig. S1. B Heatmaps of allelic contact decay profiles (CDPs) for the Xi (red bar) and Xa (blue bar) for 1161 Patski cells with at least 100 contacts per allele (along both chrX and chr1). The heatmap color scale reflects z-scaled counts within contact distance ranges shown along the x-axis. Contact counts were binned within exponentially increasing contact distances ranges (2^x where x was incremented by 0.125). These bins were then aggregated further to reduce noise by combining the counts within 10 non-overlapping bins at a time. Yellow rectangles highlight the bins used for long-range to medium-range difference (LMD) calculation (see I and J and the “Methods” section). C Plots of CDPs for the Xi (red) and Xi (blue) across the 1161 cells in B. The average CDP is plotted over the plots of CDPs for each individual cell. The Spearman correlation between the average allelic CDPs is shown at the bottom of each plot. Contact counts were binned as described in B, and the bins used for long-range to medium-range difference (LMD) calculation are highlighted in green (see I and J and the “Methods” section). D UMAP of CDPs for the Xi (red) and Xa (blue) for each of the 1161 cells in B. Contact counts were binned as described in B. The CDPs for each parental allele form two distinct clusters. E–H As in A–D, but for the B6 and spret (Sp) parental alleles of exemplar autosome, chr1. I Distributions of the difference between the long-range to mid-range differences (LMD) between the allelic CDPs for each of 1161 cells for chrX (green; as described in B) and chr1 (black; as described in F). The dashed line represents a threshold chosen below which cells were called as having an Xi bipartite structure based on the allelic CDP LMD. This threshold represents a 10% false-positive rate (FPR) based on the distribution for chr1 allelic CDP LMD (see J). J Plot of true-positive rate (TPR; proportion of cells called as having an Xi bipartite structure based on the LMD differences between allelic CDPs as described in B and I) versus false-positive rate (FPR; proportion of cells called “inactive” based on the LMD differences between the allelic CDPs for chr1 as described in F and I) at all thresholds of absolute LMD. The area under the curve (AUC) is 0.92. The dashed line represents the 10% FPR threshold chosen below which cells were called inactive based on the chrX allelic CDP LMD. K Aggregate allelic contact maps for the Xi and Xa using contact pairs from 64 mitotic cells (top row) and 2615 interphase cells (bottom row). Mitotic and interphase cells were grouped using the k-means clustering of the autosomal CDPs (Additional file 1: Fig. S2D). L Plots of allelic CDPs for the Xi (red) and Xa (blue) for cells in each of the four clusters obtained from the k-means clustering of the autosomal CDPs (Additional file 1: Fig. S2D). The average CDP is plotted over plots of the CDPs for each individual cell. The Spearman correlations between the average allelic CDPs are given at the bottom of the plot, along with cell counts used in the analysis. M–N As in K–L, but for B6 and spret alleles of chr1