As previously shown the conversion of nucleotide sequences into digital signals offers the possibility to apply signal
processing methods for the analysis of genomic data. Genomic Signal Analysis (GSA) has been used to analyze large
scale features of DNA sequences, at the scale of whole chromosomes, including both coding and non-coding regions.
The striking regularities of genomic signals reveal restrictions in the way nucleotides and pairs of nucleotides are
distributed along nucleotide sequences. Structurally, a chromosome appears to be less of a "plain text", corresponding to
certain semantic and grammar rules, but more of a "poem", satisfying additional symmetry restrictions that evoke the
"rhythm" and "rhyme". Recurrent patterns in nucleotide sequences are reflected in simple mathematical regularities
observed in genomic signals. GSA has also been used to track pathogen variability, especially concerning their resistance
to drugs. Previous work has been dedicated to the study of HIV-1, Clade F and Avian Flu. The present paper applies
GSA methodology to study Mycobacterium tuberculosis (MT) rpoB gene variability, relevant to its resistance to
antibiotics. Isolates from 50 Romanian patients have been studied both by rapid LightCycler PCR and by sequencing of a
segment of 190-250 nucleotides covering the region of interest. The variability is caused by SNPs occurring at specific
sites along the gene strand, as well as by inclusions. Because of the mentioned symmetry restrictions, the GS variations
tend to compensate. An important result is that MT can act as a vector for HIV virus, which is able to retrotranscribe its
specific genes both into human and MT genomes.