Fractal dimension approach for clustering of DNA sequences based on internucleotide distance

Abstract

Recently, the volume of biological data increases exponentially. Problem of utilization of this kind of data is not only concerning to the volume but also to its various format and storage distribution. To solve this kind of problems, some approaches require new methods, algorithms or tools to assist human being in getting beneficial from the biological data. This paper presents the usage of fractal dimension approach based on inter nucleotide distance to cluster DNA sequences. Inter nucleotide distance is a numerical representation of DNA sequences which is transformed to time series signal spectrum. Higuchi Fractal Dimension (HFD) is one of methods to estimate fractal dimension which it can be utilized to reduce time series dimension. HFD estimation then is applied to the signal spectrum and it is treated as input to clustering method. The result of this clustering shows that HFD approach can be considered as an alternative method for dimensional reduction purposes. Compared with previous study result as ground truth, the HFD approach clustering provides some similarities in certain degree. Tested with two kinds of data test sample, this approach results 6 and 7 group similarities of 10 groups.