Extended Data Figure 5: Allele-specific chromatin structure.

a, Validation of haplotypes by (i) RNA-sequencing (i) and whole-genome sequencing (WGS) (ii). Shown in dark blue is the percentage of reads connecting variants in the same predicted haplotype, while in light blue is the percentage of reads connecting variants predicted to be on different haplotypes. b, Inset labelled with an asterisk is from Fig. 3b showing DHS sequencing reads over a SNP upstream from the SNRPN gene, demonstrating how different chromatin features can be assigned to a given haplotype. c, Scatter plots showing the correlation coefficient of PC1 values obtained from compartment A/B analysis between the two biological replicates for each allele. Despite the reduction in reads when Hi-C data are split into two alleles, the PC1 scores were highly reproducible between replicates. d, Shown is a genome browser image of PC1 values in chromosome 2 for the p1 allele, p2 allele, and for all Hi-C reads without resolving the two alleles. PC1 scores are highly consistent, suggesting that homologous chromosomes fold in highly similar patterns. e, Scatter plot of PC1 values between the p1 and p2 alleles in H1 and H1-derived lineages. The Pearson correlation coefficient value is 0.96. f, Fraction of the genome that shows changes in A/B compartment status across alleles. For ES, ME, MS, NP and TB cells, 1.3%, 0.6%, 1.9%, 2.3% and 1.6% of total genomic regions shows allelic compartment A/B patterns, respectively. g, Percentage of allele-biased (purple) or non-allele-biased (grey) genes that have different A/B compartment status in each lineage. Only in ES cells is there a significant association between allele-biased genes and regions with variable A/B compartment between alleles (Fisher’s exact test). h, Similar to g, but showing the association between imprinted genes and changes in A/B compartment between alleles. No lineage shows a significant association.