Fig. 5
Validation of the predicted CRMs by VISTA enhancers, FANTOM promoters (FPs), FANTOM enhancers (FEs) and QTLs. A Sensitivity (recall rate) of the predicted CRMs or the control sequences as a function of p-value cutoff for recalling each set of the experimentally determined elements. The dashed vertical line indicates the p-value cutoff of 0.05. The sensitivity of the CRMs predicted at all the indicated p-value cutoffs are significantly higher (p < 2.2 × 10–302, χ2 test) than the control sequences for recalling each set of the experimentally determined elements. B Distributions of phyloP scores of the shared and unshared nucleotide positions of the elements in each set of the experimentally determined elements, in comparison with those of the predicted CRMs at p ≤ 0.05 and of the non-CRMCs. The difference between the distributions of shared and unshared positions in each set of the experimentally determined elements is significant, p < 2.2 × 10–302 (K-S test). Note that there are only three unrecalled VISTA enhancers