I'm writing a program to find the promoter regions in bacteria. I though that the simple idea of getting let's say 100 nucleotides before the coding sequence on the DNA and calculating the consensus would work (at least would show some pattern). But, apparently, it doesn't.
Let's say we've got E.coli genome and annotations from NCBI RefSeq. Next we scan through annotation to get positions of CDS and record 100 nucleotides before the CDS (actually, I include first 3 nucleotides of CDS, usually ATG). https://ibb.co/kErQfw
But I see nothing similar to the Pribnow box. What am I doing wrong?