Long Range Correlations Between Nucleotide Triplets

Abstract

The study of statistical fluctuations in DNA sequences can reveal important characteristics of their organization. Particularly, in the last two decades, several studies have focused on the detection of Long Range Correlations (LRC) in DNA sequences, and on the study of particular long range dependence properties of nucleotides. Since protein coding is carried out by codons (nucleotide triplets), we conduct LRD analysis to see whether there is a long range dependence among nucleotide triplets. Our interest is not limited to auto correlations of single codons, but also extends to cross correlations over all possible pairs of nucleotide triplets. The analysis on nucleotide triplets reveal LRCs for the human Chromosomes 20 and 21. Moreover, some triplets that contain only nucleotides Adenine and Thymine are seen to exhibit correlation significantly higher than others.