Wavelet Applied to the Classification of Bacterial Genomes

Abstract

The classifications resulting from phylogenetic analysis are essential tools for evolutionary studies. Phylogenetic is more than a part of evolutionary biology because its underlying philosophy provides a way to see nature, ask questions, and solve problems related to the evolution of organisms. Given the importance of phylogeny, our aim was to devise a method to assess the delimitation of bacterial species. We used the non-decimated discrete wavelet transform. The wavelet function used was Daubechies' with four null moments, considering seven, four and two decomposition levels. For clustering, the energy (variance) obtained at each level of decomposition and the Mahalanobis distance was used to visualize the dendrogram formation process. Through the analysis, we verified that the gram-positive bacteria were classified well into their respective species, but most gram-negative bacteria did not take into account the more significant amount of energy obtained in scenario two. According to the results, the energy plays an important role in the delimitation of groups of bacterial species.