Specific genetic deletion are identified in mycobacteria isolates, including variations in the M. tuberculosis genome sequence between isolates, and numerous deletion present in BCG as compared to M. tb. These deletions are used as markers to distinguish between pathogenic and avirulent strains, and as a marker for particular M. tb isolates. Deletions specific to vaccine strains of BCG are useful in determining whether a positive tuberculin skin test is indicative of actual tuberculosis infection. The deleted sequences may be re-introduced into BCG to improve the efficacy of vaccination. Alternatively, the genetic sequence that corresponds to the deletion(s) are deleted from M. bovis or M. tubercuosis to attenuate the pathogenic bacteria.