DNA encodes multiple independent levels of functional information. For instance, one level encodes proteins and sequence targets for DNA-binding factors, another one is contained in the physical and structural properties of the DNA molecule itself, others are the signals that guide higher degrees of genome organization. In this sense DNA can be considered as a four-symbol language text, endowed with semantic meaning and a precise syntax. In order to crack these codes, lots of methods have been used, including the ones that derive from probability theory, specific entropy indicators, measurements of long- and short-range correlations and various signal-processing models.
In our talk we are going to deal with gene promoters, regulatory regions belonging to the vast noncoding portion of the genome, where the chemo-physical properties of the sequence of nucleotides are crucial to determine the double helix dynamics and, accordingly, the transmission of the genetic message. First, we are going to show our entropic analyses of base composition, periodicity and information content which point out the relation between specific motifs along the sequence and their role in determining regulatory functions. Then, we are going to present the results obtained by combining two spectral methods: one clustering algorithm, based on an alignment procedure, the other one based on a standard spectral analysis of the eigenmodes of the nonlinear chains representing promoters according to the Peyrard-Bishop DNA model. Through these methods, it is possible to recover the regular sequences that characterize different functional classes of promoters and to find out elements with important biological meanings.