Fig. 2
From: Global chromatin conformation differences in the Drosophila dosage compensated chromosome X
More dispersed cis-interactions and more trans-contacts in male chrX. a Cartoon to illustrate the selection of top-scoring interactions. In the normalized Hi-C contact map, for each diagonal (red lines), the data points with highest signal along the entire chromosome (top 5%, i.e. higher than 95th percentile, with default settings) are selected (yellow squares). b Representative (3 Mb) regions from contact matrices for an autosome arm (chr3R, left plot) and chrX (right plot). In each composite matrix, the top-scoring interactions (top 5%) are marked for male (blue points, upper triangle) and female (red points, lower triangle) sex-sorted Hi-C data at 25 kb bins resolution and normalized with chromosome-wise ICE. c The difference in the fraction of clustered top-scoring interactions (female–male) is shown for each autosome arm and chrX, based on 25 kb binned chromosome-wise normalized Hi-C data for sex-sorted embryos. Clustered top-scoring interactions were defined by aggregating neighbouring top-scoring data points (see Methods) and considering only distances up to 2 Mb. d The relative amount of trans-contacts for each chromosome is reported as the ratio of the number of trans- over cis-mapping read pairs for each chromosome in male and female sex-sorted embryos datasets. e The propensity of each chromosome to participate in trans-contacts is shown for both male (left) and female (right) samples. The trans-mapping read pairs originating from each chromosome (rows) are grouped based on the target chromosome (columns) and their number is compared (log2 ratio) with the random expectation. In the random expectation model, trans-reads originating from any chromosome are expected to be uniformly distributed over the other (target) chromosomes, with their relative distribution depending on the size (and copy number) of each specific target chromosome (see Methods). The log2 ratio between the observed and expected fraction of trans-read counts is reported as a colour gradient. Note that the heatmap is not expected to be symmetrical because the expected number of interactions is different depending on the respective size of the origin vs target chromosome in each pair. The diagonal is grey as cis-interactions are not considered.
