Fig. 2

Comparison of experimental functional assays with in silico ncER predictions. a CRISPRi effect on cell viability (75,622 sgRNA probes pairs) from Fulco et al.³² and the corresponding maximum ncER score across the GATA1 and MYC loci. Accuracy at four ncER thresholds is shown in yellow, orange, red and dark-red respectively for the 95th, 99th, 99.5th and 99.9th percentiles. b Distribution of maximum ncER at different bins of cell viability (0 to less than −3 log2 fold change). The boxplot’s central line represents the median, the bounds represent the 25th and 75th percentile, and the whiskers extend up to 1.5 the interquatile range from the respective bounds. p Values were computed with independent two-group Mann–Whitney unpaired test. c Fraction of regions with at least one high ncER score, as defined by four different ncER percentile thresholds. CRESTseq peaks are shown in red (N = 45), random matched sized in silico regions extracted from the same locus (“random same locus, N = 4500) in dark gray and random matched sized in silico regions extracted genome-wide (“random GW”, N = 4500), in light gray. p Values were computed with Fisher's exact test. FC fold change, CRESTseq cis-regulatory elements by tiling-deletion and sequencing, GW genome-wide, CRISPRi clustered regularly interspaced short palindromic repeats interference