CDS identification in a viral single-strand DNA (ssDNA) using Fisher linear discriminant

In this paper, a novel discriminant algorithm based on Fisher linear discriminant that enables distinguishing the features of Codon DNA Segments (CDS) in a DNA sequence consistent with the underlying stochastic attributes is developed. It is specifically applied to a viral single-strand DNA (ssDNA). The Fisher discriminant adopted delineates judiciously, the codon and non-codon regions in the test ssDNA. With a complete description of the computational procedure, the efficacy of the algorithm is tested via simulated experiments on a real ssDNA of B19 virus. The results are validated against GenBank data available on the test viral sequence. Inferences with regard to CDS regions present in the test sequence are made indicating possible fuzzy transitions of codon/non-codon sections.