Fig. 4: Genomic windows enriched for Hi-C- and GAM-specific contacts are clustered in the genome.

a, Strategy for identifying genomic regions forming many contacts specific to either GAM or Hi-C. We counted how often a genomic region was an anchor point in the set of GAM-specific contacts, Hi-C-specific contacts or strong-and-common contacts. The 10% of genomic windows with the highest absolute difference between the number of GAM-specific and Hi-C-specific contacts were classified as Hi-C-preferred regions or GAM-preferred regions, respectively. Similarly, the top 10% of strong-and-common contacts were used to define common regions that participate to a similar extent in strong contacts seen in both methods. b, Example region on chromosome 4 showing the delta z-score matrix (top) and clusters of method-specific preferences designated as Hi-C-preferred regions, GAM-preferred regions or common regions (bottom). c–f, Characteristics of Hi-C-preferred, common and GAM-preferred windows. c, Total number of genes present in windows of each category. d, Transcriptional activity (log₂-scaled transcripts per million per gene, TPM) of all genes per category. Bars mark the 25%, 50% and 75% quantiles. e, Proportion of windows in lamina-associated domains (LADs) for each category. Numbers on top of the bars are the number of LAD windows and the total number of windows for each category. f, Numbers of genomic windows in each category that overlap a range of features or are classified as compartment A or B.