Fig. 4

Correlations between REDfly enhancers overlapping enhancers from each EnhancerAtlas subset used for the “L3_wing_disc” EnhancerAtlas set. The correlation structure demonstrates that experimental batch effects predominate over assay-type effects. High correlations are seen between sets from the same experiment group (see white box, yellow box), even when assay types differ (dotted yellow box). Although assay-specific correlations are sometimes present (yellow asterisks), identical assays can also be poorly correlated when from different experiment groups (white asterisks). Individual experiments on the y-axis (“GSM” identifiers) are colored according to common experimental series (i.e., performed by the same laboratory as a specific set of experiments) as provided in GEO (“GSE” identifiers, see Table S1d [Additional File 1]). Sets labeled in black are unrelated. Label colors on the x-axis indicate assay types as follows: red, H3K4me1 ChIP-seq; green, H3K27ac ChIP-seq; orange, Grh ChIP-seq; cyan, Pol2 ChIP-seq; dark blue, ATAC-seq; black, ChIP-seq against various transcription factors