Fig. 4.

Super-enhancers and HIV-1 occupy the same 3D sub-compartment. a Definition and identification of 3D compartments. For each chromosome, 15 spatial communities were identified by clustering. The inter-chromosomal contacts between the communities were used as a basis for another round of clustering in five genome-wide spatial communities. b Pie chart showing the coverage of the sub-compartments in the Jurkat genome. c Distribution of AB scores in the 3D sub-compartments. The AB score measures the likelihood that a locus belongs to the A or B compartment. Extreme values +100 and −100 stands for “fully in A” or “fully in B”, respectively. A score of 0 means “both or neither.” d Proportion of 3D sub-compartments covered by major chromatin features. Coverage was computed as the span of enriched ChIP-Seq signal divided by the sub-compartment size. e Spie chart showing the observed vs expected HIV-1 insertions in the sub-compartments. The expected amount of insertions is the area of the wedge delimited by the circle in bold line, and the observed amount is the area of the colored wedge. Dotted lines represent the limit of the wedge for 2× and 3× enrichment outside the circle, and 0.5× depletion inside the circle. The observed/expected ratio is approximately 2.5 times higher in A1 than in A2. f Box plot showing the expression of protein-coding genes in the sub-compartments. The plot was rendered using defaults from the ggplot2 library. The y-axis has a logarithmic scale. g Bar plot showing the integration density of HIV in super-enhancers located in different 3D sub-compartments. h Bar plot showing the contribution of different predictors to the HIV-1 insertion sites in typical genes (left) or in hotspots (right). The y-axis represents the loss of accuracy when the corresponding variable is removed from the model. Expr. gene expression, dSE distance to nearest super-enhancer, Size gene size, 3D sub-compartment. See “Methods” for detail