Entropies and word frequencies in the Eukaryotic Promoter Database

The comprehension of the mechanism of gene expression is one of the more relevant fields associated with genome analysis. The experimental sequencing activities evidence the existence of control domains like promoter, enhancer and silencer. The promoter signals are well documented in various databases. Inside these regions are distributed a number of characteristic sequences involved in the interactions between the DNA and the transcriptional complex. Different approaches have been applied in order to detect these features in genomic sequences. In the present paper we investigate the statistical properties of a subset of the Eukaryotic Promoter Database (EPD using entropies as a measure of complexity. An efficient computational approach, suitable for a coarse-grain parallelization, is used for a fast processing of the symbolic sequence. Results show a large deviation from the randomness for the distribution of words in the EPD database.