Figure 2. DNA FISH confirms high pairing frequency for the KvDMR region.

A) Examples of paired DNA FISH spots in E13.5 foetal liver. Red: probe covering the KvDMR region, green: probe covering the region around the myc gene, blue: DAPI counterstain. B) Pairing frequency of the KvDMR and myc regions. Each dot represents one biological sample, and for each sample 300 nuclei were counted in four technical replicates. Differences were assessed by unpaired t-test, ns: not significant, *: p<0.05, ***: p<0.001. KvDMR signals show higher pairing frequency than myc signals. C) Frequency distributions of radial distances of unpaired (black) and paired (blue) KvDMR DNA FISH signals in E13.5 liver (n = 93). Radial distances >1 can occur if the nucleus is not a perfect sphere. Radial distances are not different between paired and unpaired KvDMR FISH signals (t-test, p = 0.0668) indicating that pairing events can happen at all nuclear locations KvDMR alleles normally occupy. D) Distances between homologous alleles in E13.5 liver represented as Tukey box-whisker plots (n = 600). Interallelic distances were normalised to the radius of the nucleus (distance/radius). Differences were assessed by one-way ANOVA with Bonferroni’s multiple comparison post test. Simulated: a group of spots displaying the same radial distributions as KvDMR and myc FISH signals, respectively, were placed into a sphere at random and their interallelic distances determined (see Fig. S1A, B and Methods). While interallelic distances between myc signals show a distribution expected from their radial positions (no difference between sample and simulated, p>0.05), KvDMR signals are significantly closer together than expected (difference between sample and simulated, p<0.001). Also, KvDMR signals are generally closer together than myc signals. The same data is represented as a histogram in Fig. S1C to illustrate the presence of a subpopulation of KvDMR signals that display very short interallelic distances.