Fig. 3

cDNA contamination in published datasets. A Examples of ATAC-seq signal from two primary tumor samples from the TCGA [5] showing contamination with DDX58 (encoding the antiviral innate immune response receptor RIG-I; left) or the cohesin component STAG2 (right). True signal would be expected at the promoter and potential intragenic regulatory elements, but not over all exons. Arrowheads indicate spurious signal peak calls caused by contaminant reads over exons (black boxes in gene track; peak calls obtained from ref [5]). B cDNA contamination with PPARG in a FOXK2 ChIP-seq experiment in HEK293T cells and PAX7 in an EZH2 ChIP-seq experiment in HUVEC cells from the ENCODE project. Arrowheads indicate official ENCODE peak calls due to contaminant signal over exons. C Examples of cDNA contamination with prostate cancer genes FOXO1 and SPOP in an androgen receptor (AR) ChIP-seq experiment performed in the prostate cancer cell line C2-4 [36, 37]. D Example of transduced human NRAS cDNA in a mouse ATAC-seq experiment in cell line ICC2.7