Sequence bias in chromatin fragmentation leads to misinterpretation of protein-DNA interactions in vivo

Genome regulation depends on the binding of regulatory proteins to chromatin. V-plots are widely used to assess the binding of transcription factors and other protein complexes to DNA in vivo. However, we show here that V-plots result from the cleavage sequence preferences of the chromatin fragmentation methods and can be generated equally well on naked deproteinised DNA as on chromatin, and on any randomly selected sequence motif in the genome. Our results argue against using chromatin fragmentation methods, combined to previous knowledge of motifs as the only method to map DNA-associated proteins by stacking the end of sequence reads over genomic regions that share a common nucleotide motif.