Dear Prit,

I am not a specialist, but when it comes to finding a motif in an assembly I usually run this tool:

http://ccg.vital-it.ch/pwmscan/

You can generate a position weight matrix with the E-box motif and try it.

You will get a number (for E-box probably a lot) or regions with coordinates you can overlap with UCSC tables (Refgene, mirBase...).

Otherwise have you tried to mine the ENCODE datasheet for this (using Myc binding sites for example)?

Good luck,

Antoine.

An E-box is a pretty non-specific motif and it tends to pop up a lot even in areas not known to bind a specific transcription factor. In its most general form it only has 4 specific positions and is a palindrome. Many E-box motifs for specific TFs are more specific with one or both of the middle positions defined.

A general motif scanner looking for the pattern CANNTG is likely to find a lot of false positives, simply because it is such a low-information motif.

If you want to try that I can help you. It is fairly easy technically. But as mentioned by previous answers, the result is going to be a mess. I have done a analysis that enriched all TF binding sequences in promoters. The most over-represented sequences are always the short ones. You have to perform a statistical analysis to show that the over-representation of a TF binding sequence is not by random.