Figure 2.

Proximity-ligation products are predominantly intrahaplotype. (a) Heat map of whole-genome, long-range, chromatin contact frequencies. Hi-C reads originating from the CAST (“c”) or J129 (“j”) genome were distinguished based on the known haplotype structures of the parental strains. The frequency of interactions between each allele of each chromosome was calculated using 10-Mb bin size. The CAST or J129 allele of each chromosome primarily interacts in cis, confirming that the chromosomes territories seen in Hi-C data occur for individual alleles. Inset shows a magnified view of the CAST and J129 alleles for chromosomes 12 through 16. (b) Chart of intrahaplotype (cis) and interhaplotype (h-trans) interaction frequencies. From a priori haplotype information, we distinguish Hi-C read-pairs as interacting in cis (green) and in h-trans (purple). Plot was generated using data from chromosomes 1–19 in the CAST×J129 system. In (i), we used all intrachromosomal reads, and in (ii), we excluded all intrachromosomal reads that map with an insert size <1kb, as these are probably short contiguous DNA fragments and are therefore very likely to be in cis. Thus, the analysis described in (ii) provides a more conservative estimate of h-trans. Comparing these charts, the frequency of h-trans is at most ~2%. (c) Comparison of the h-trans interaction probability as a function of insert sizes. Plot generated using data from chromosomes 1–19 in CAST×J129 system. LOWESS fit (purple) was performed at 2% smoothing. Below 30 Mb, the probability of a read being an h-trans interaction is ≤5% (green dashed line) at any given insert size. (d) Similar to b, but excluding reads that have inserts >30 megabases. Probability of h-trans is estimated to be ~5% at 30-Mb insert size. Therefore, we use this cutoff as a maximum insert size for subsequent analyses.