Statistical algorithms for long DNA sequences: Oligonucleotide distributions and homogeneity maps

Article type: Research Article

Authors: Katsaloulis, P.^; | Theoharis, T. | Provata, A.

Affiliations: Department of Informatics, University of Athens, 15784 Athens, Greece | Institute of Physical Chemistry, National Research Centre "Demokritos", 15310 Athens, Greece

Note: [] Corresponding author. Tel.: +30 210 7275106; E-mail: theotheo@di.uoa.gr

Abstract: The statistical properties of oligonucleotide appearances within long DNA sequences often reveal useful characteristics of the corresponding DNA areas. Two algorithms to statistically analyze oligonucleotide appearances within long DNA sequences in genome banks are presented. The first algorithm determines statistical indices for arbitrary length oligonucleotides within arbitrary length DNA sequences. The critical exponent μ of the distance distribution between consecutive occurrences of the same oligonucleotide is calculated and its value is shown to characterize the functionality of the oligonucleotide. The second algorithm searches for areas with variable homogeneity, based on the density of oligonucleotides. The two algorithms have been applied to representative eucaryotes (the animal Mus musculusand the plant Arabidopsis thaliana) and interesting results were obtained, confirmed by biological observations. All programs are open source and publicly available on our web site.

Journal: Scientific Programming, vol. 13, no. 3, pp. 177-188, 2005