GMM-Based Classification of Genomic Sequences

At present many digital signal processing based techniques are available to predict genomic protein coding regions. However, accurate identification of these regions at the level of individual nucleotides remains a challenge. In this paper, we propose the novel use of a multi-dimensional feature and Gaussian mixture models for the classification between protein coding and non-coding nucleotides. Employing signal processing based time-domain and frequency-domain features, the novel system described herein is shown to produce identification accuracies of more than 75% and 79% respectively for protein coding and non-coding nucleotides, when evaluated on the GENSCAN data set.