Contents I. Introduction
Graphical representations of DNA offer visual inspection of DNA sequences [1]. However, in [2], the author investigated corrections that reveal some aspects of similarity which could not be determined through the traditional alignment-based methods. The space of similarity for complex objects is multidimensional. Complex objects may be similar in one aspect; however, it can be very different in another one. Recently, many numerical characterizations for DNA or protein sequences have been introduced, where most of numerical characterizations are extracted from the string representations and graphical representations. The simpler and more important feature from string representations first used for comparison of genome sequence [3] and later for alignment-free comparison of regulatory sequences [4]. Various frequency-based algorithms later have been introduced for sequence comparisons, as indicated in [5] and [6]. Besides the representations based on single nucleotide, the dinucleotide analysis has also been tried by several authors. Randić [7] proposed a condensed representation of DNA based on pairs of nucleotides. Wu et al. [8] proposed the analysis approaches based on neighboring nucleotides of DNA sequence, which reveal the biology information hidden between dual nucleotides.
