Briefly, raw sequencing data were de-multiplexed and aligned using either the Mouse July 2007 (NCBI37/mm9) or Zebrafish December 2008 (Zv8/danRer6) assembly as the reference genome. Reads located in fragments flanked by two restriction sites of the same enzyme, or in fragments smaller than 40 bp were filtered out. Mapped reads were then converted to reads-per-1kb-bin units, and smoothed using a running mean 5-fragment window algorithm. To calculate the statistically significant targets for each viewpoint, a background theoretical model was calculated as the exponential fit curve of the average signal of 61 4C-Seq different samples33–35. The p-value for each bin was calculated by means of Poisson probability function. Smoothed data were uploaded to the UCSC Browser36 for visualization and plotted using the browser and Circos37. For plotting, statistically significant targets were defined as p<0.01. Data was submitted to GEO under accession GSE52830, and can be accessed online at http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=elcraaeglhstzwz&acc=GSE52830.
