Box counting method in recording, processing and evaluation of genomic signals

Fractals are geometrical shapes with noninteger dimension andinvariance against change of scale factor. For each geometricalshape, a single parameter - dimension - can be calculated. Forcalculation, Box Counting Method (BCM) was chosen.Determination of dimension on one level of resolution is notsufficient, it is necessary to subsequently process it and determine multifractal coefficient. Aim of our interest consists in analysis of images obtained from a linear sequence of DNA code. For analysis itself, sequences of Homo sapiens and Citrobacter youngae were chosen. Results of calculation are fractal coefficients derived from dimensions of generated structures. This result enables to introduce criterion for sequences determination. Graphical outputs may bealso represented as multidimensional alternative transformation of linearly recorded genomic signal. Algorithms were developed in computational environment MATLAB. Data were downloaded from a public database EMBL (ESI) and GenBank (NCBI).