Genomic signature is preserved in short DNA fragments

The "chaos game representation" (CGR) paradigm has been implemented to display the use of short oligonucleotides in genomes in the form of fractal images. These images can be considered as a genomic signature. Using an unsupervised classification approach, it is shown that short fragments of genomic sequences retain most of the characteristics of the species they come from. It thus appears possible to perform a global comparison of species by means of genome fragments found in databases. The efficiency of this approach is evaluated as a function of the size of the fragments and the length of the oligonucleotides.