4. Discussion
The number of bacterial genomes stored in the GenBank database increases every day. Analysis of this wealth of information creates a greater knowledge into genetics, biotechnology and health-related that are important to the economic aspects, research and a number of diseases [19].
Genome alignment has been a very important method to understanding evolution, gene function and phylogenomic comparison. The recognition sites with high probability of hybridization of each bacteria give us a genomic fingerprint unique and specific to each microorganism, this compared with MuMmer 3.0 to identify changes in the genome.
The analysis of eight Bacillus anthracis genomes with the construction of two phylogenomic trees reveals the high similarity between these species; this was confirmed by the analysis conducted using the MUMmer 3.0. This software compares two bacterial genomes by representing them in a dot plot; the plot shows the sections of genome that are shared and denotes the similarity between the two genomic sequences.
A range of methods is available for identifying bacteria and their choice depends on the techniques and resources available. In this study we classified Bacillus anthracis with basis on their genomic fingerprint, which is unique and specific for each strain. Our results will help to confirm and compare with results obtained with other, traditional methods for classification and identification, using the Universal Fingerprint Chip. 
Microarray hybridization has been implemented for bacteria identification using sequences related to the 16S rRNA unit as target, and a collection of 14,283 probes, each 12 or 13 nucleotides long, has been designed. However, this approach fails to discriminate between closely related species due to the high similarity of their sequences [20]. In a previous study by Bavykin, proposed a set of probes, based on the 16S rRNA unit, for the detection of various bacteria-related diseases and to identify potential binding sites based on the microarray technique [21]. Classification of different organisms and strains with basis on the 16S rRNA unit is not quite reliable, as those studies have shown a high similarity between rRNA sequences and their inability to discriminate between highly similar strains [7]. However, some in situ hybridization has been demonstrated capable of discriminating between highly similar strains such as Bacillus anthracis [22].
The extended method yielded more reliable results than the direct method because, in the former, the formation of the highly potential sites is not at random as this method adds four nucleotides to both ends of the 13-nucleotide probe to yield a 21-nucleotide long (4 + 13 + 4) segment. This makes it more difficult to find, by chance, a similar 21-nucleotide long probe. Each of the sites thus identified confirm that the duplex formation and potential sites are not random [14].
Nevertheless, comparing genomes by means of their genomic fingerprint is interesting because: (a) UFC probes distribute themselves randomly and uniformly throughout the genome under study, so that the fingerprint is representative of the genomic sequence; (b) comparison does not require prior alignment of the genomes; and (c) tools such as the extension technique help to determine whether the regions compared between genomes have a common evolutionary origin (homology) [23].
The number of virtual hybridization signals obtained with the UFC-13 has a very close relationship, but with significant variation, with the size of the genome of the organisms. Such variation can be accounted for by between-sequences differences and by differences in the genome G + C content [24].