Fig. 2: GAM and Hi-C show strong, method-specific contacts.
From: Multiplex-GAM: genome-wide identification of chromatin contacts yields insights overlooked by Hi-C
a, Strategy to assess differences and similarities between GAM and Hi-C contact maps (chr. 4, 80–90 Mb). GAM and Hi-C contact data have different distributions (1); therefore, contacts at the same genomic distance undergo z-score transformation (2). Hi-C z-scores are then subtracted from GAM z-scores to generate a delta z-score matrix (3), from which we extract the 5% most differential contacts between GAM and Hi-C (GAM-specific or Hi-C-specific; 4) and the top 10% of contacts common for both methods (strong-and-common; 5). b, Interactions identified by SLICE or by Fit-Hi-C shown in the distribution of delta z-scores. c, Overlap of contacts co-occurring in combinations of the top 20% strongest contacts from GAM and Hi-C, and the 10% strong-and-common set. Intersection groups are colored by fraction supported by SLICE or Fit-Hi-C.
