Multiple novel promoter-architectures revealed by decoding the hidden heterogeneity within the genome. Nucleic Acids Res

An important question in biology is how differ-ent promoter-architectures contribute to the diver-sity in regulation of transcription initiation. A step forward has been the production of genome-wide maps of transcription start sites (TSSs) using high-throughput sequencing. However, the subsequent step of characterizing promoters and their func-tions is still largely done on the basis of previ-ously established promoter-elements like the TATA-box in eukaryotes or the-10 box in bacteria. Un-fortunately, a majority of promoters and their activ-ities cannot be explained by these few elements. Traditional motif discovery methods that identify novel elements also fail here, because TSS neigh-borhoods are often highly heterogeneous contain-ing no overrepresented motif. We present a new, organism-independent method that explicitly mod-els this heterogeneity while unraveling different promoter-architectures. For example, in five bacte-ria, we detect the presence of a pyrimidine preced-ing the TSS under very specific circumstances. In tu-berculosis, we show for the first time that the spac-ing between the bacterial 10-motif and TSS is uti-lized by the pathogen for dynamic gene-regulation. In eukaryotes, we identify several new elements that are important for development. Identified promoter-architectures show differential patterns of evolution, chromatin structure and TSS spread, suggesting dis-tinct regulatory functions. This work highlights the importance of characterizing heterogeneity within high-throughput genomic data rather than analyzing average patterns of nucleotide composition