We considered constructing three 8-component vectors for a DNA primary sequence using triplets of nucleic acid bases. For two DNA sequences, using the corresponding vectors, we constructed a set of 3 x 3 matrices called the related matrix. The normalized leading eigenvalues from the constructed matrices were selected as invariants to characterize the degree of similarity between the two DNA sequences. Construction of similarity/dissirnilarity tables based on this invariant for sequences of DNA of the first exon of the P-globin gene from I I species illustrates the utility of the newly formulated matrices for DNA. (c) 2005 Elsevier B.V. All rights reserved.