B-Splines for Genomic Signatures

A variety of data analysis concerned with genome sequences support the proposal that each living organism owns a genomic signature. Classical approaches to the genomic signature deciphering are based on the counting of oligonucleotides in DNA sequences. Statistical tools as well as graphical representations have been used to associate genomic signatures with oligonucleotides frequencies in the genome. Moreover, many experimental results also show that the core features characterizing the whole genome are preserved in short subsequences, validating the idea of a genomic signature. In this paper we present a new method to detect the genomic signature. The proposal stands on two ideas: first, give a simple mathematical representation of trinucleotide frequencies; second, deal it with B-spline functions. The result is a continuous function representing the genomic signature. The advantages of our method resides in maintaining the essential properties of the signature itself and providing a more formal instrument of analysis.