Analyzing DNA Sequences Using Clustering Algorithm

Data mining gives a bright prospective in DNA sequences analysis through its concepts and techniques. This study carries out exploratory data analysis method to cluster DNA sequences.Feature vectors have been developed to map the DNA sequences to a twelve-dimensional vector in the space. Lysozyme, Myoglobin and Rhodopsin protein families have been tested in this space. The results of DNA sequences comparison among homologous sequences give close distances between their characterization vectors which are easily distinguishable from non-homologous in experiment it with a fixed DNA sequence size that does not exceed the maximum length of the shortest DNA sequence. Global comparison for multiple DNA sequences simultaneously presented in the genomic space is the main advantage of this work by applying direct comparison of the corresponding characteristic vectors distances. The novelty of this work is that for the new DNA sequence, there is no need to compare the new DNA sequence with the whole DNA sequences length, just the comparison focused on a fixed number of all the sequences in a way that does not exceed the maximum length of the new DNA sequence. In other words, parts of the DNA sequence can identify the functionality of the DNA sequence, and make it clustered with its family members.