2.7 Boundary correlation experiments
Boundary correlation was performed as previously reported (Dixon et al., 2012). We merged the center position of the boundary between two experiments of interest and calculated the directionality indexes ± 10 bins for each center. Then we randomly selected 20 bins from each of the two cell lines and calculated the spearman correlation between the two vectors. We attained the Spearman correlation coefficients distribution after repeating the randomization 10,000 times. We defined values at the top 1% of distribution as significant correlation threshold. Boundaries were identified as ‘cell type specific’ if the boundary region was in one cell and lacked a significant correlation between the LDC and WCC liver cells.