Symbolic to numerical conversion of DNA sequences using finite-context models

Symbolic sequences can be analysed using two main approaches. One is by means of algorithms specifically designed for processing symbolic sequences. The other uses signal processing techniques, after converting the sequence from symbols to numbers. The latter approach depends on the availability of meaningful numerical representations of the sequences. In this paper, we present a technique that uses finite-context models to generate numerical information sequences from the DNA symbolic data. We give some examples that illustrate the method and show that these information sequences may reveal important structural properties of the DNA sequences. Moreover, the proposed approach is fast, allowing a quick bird's-eye view of whole chromosomes, with the aim of locating potentially interesting regions.