Neural Network in Human Identification by DNA Sequences

In this paper we propose a new method to analyze the similarity/dissimilarity of DNA sequences which can be used in human identification field. This method is based on the graphical representation proposed by Randic et al [M. Randic, M. Vracko, L. Nella, P. Dejan, Chem. (2003)]. Instead of calculating the leading eigenvalues of the matrix for graphical representation we smooth the zigzag curve and calculate its curvature. Similarity between DNA sequences are decided by neural network. Our method is useful for human identification in criminal investigations and in genetic disease. Our results verify the validity of our method.