We selected enhancers for validation based on human-zebrafish conservation (>70% sequence identity over 100 nt, hg19 vs DanRer7) in order to take into account the large evolutionary separation between the two species, and selected enhancers that were only expressed in a subset of tissues/cells. We did not take epigenetic data (ChIP/DHS etc.) into consideration. We also selected three negative control regions, chosen randomly from the human genome with the following constraints: low conservation with zebrafish and no other enhancer-selective feature, that is, no DNase hypersensitivity, no H3K4me1 or H3K27ac signals and CAGE signal only at noise levels.
