Fig. 3

Binding characteristics: qfrag length distribution with pseudo-control. a The 5’-3’ (λ) exonuclease is employed to trim the DNA sequences of a fragment of length ℓ to within a few bp at which the protein of interest has been bound to the DNA. This yields shorter fragments of length ℓ ^′ and ℓ ^″. The width of the protected region (ℓ ^‴) is given by the distance between two 5’ ends on the forward and reverse strand. b Schematic representation of the pseudo-control and the corresponding transformation. For mapping artifacts predominantly to be found on chrU and chrUextra (left-hand) and for genuine ChIP-nexus peaks (right-hand). The pseudo-control is derived from the original mapping data by swapping the strand of each read and subsequently shifting the 5’ end by one read length towards 5’ direction. For artifacts, this has no effect on the qfrag-length distribution in the pseudo-control. c qfrag-length distribution for original data (black) and pseudo-control (gray). Both distributions are are dominated by the phantom peak at one read length. d We use the difference between the qfrag-length distributions as signature and the maximum at a length of 19 as estimate for ℓ ^′′′