Local scaling and multifractal spectrum analyses of DNA sequences – GenBank data analysis

Base sequences of deoxyribonucleic acid (DNA) in an organism carry all the instructions regarding its growth and development. On the surface, such sequences seem irregular; yet in reality, they are symbolic sequences with an organized structure. This study investigates the characteristics of base arrangement and distribution in DNA sequences from the fractal theory viewpoint. In addition to multifractal features demonstrated by the DNA sequence, this study also compares the multifractal spectra derived from a particular family of gene among several different species. The results reveal that a considerable correlation exists between base distribution and evolutionary order. Furthermore, local scaling exponent (Hölder exponent) differences between coding segments (exon) and non-coding segments (intron) are also examined. It is suggested that such differences in the local distribution of bases can be applied to find coding segments within the DNA sequence that is to be translated into protein. This local scaling analysis is feasible and has the potential to become an effective tool for rapid location of possible coding sites in DNA sequences. The authors hope that future studies using more complicated bioinformatics methods for analyzing DNA sequences can benefit from this study.