Experimental versus in silico fluorescent amplified fragment length polymorphism analysis of Mycobacterium tuberculosis: improved typing with an extended fragment range.

Whole-genome fingerprinting fluorescent amplified fragment length polymorphism (FAFLP) data were compared with in silico data for the sequenced strains of Mycobacterium tuberculosis (H37Rv and CDC1551). For this G+C-rich genome, many predicted fragments were not detected experimentally. For H37Rv, only 108 (66%) of the 163 predicted EcoRI-MseI fragments between 100 and 500 bp were visualized in vitro. FAFLP was also used to identify polymorphism in 10 clinical isolates of M. tuberculosis characterized previously by IS6110 typing, examining fragments of up to 1,000 bp in size rather than up to 500 bp as was done previously. Five isolates had unique IS6110 profiles and were not known to be epidemiologically related, two isolates were the same single-band IS6110 type but were not known to be epidemiologically related, and the remaining three isolates were epidemiologically related with identical IS6110 profiles. Analysis of fragments in the 500- to 1,000-bp range using nonselective primers differentiated better between strains than analysis of fragments in the 50- to 500-bp range using a set of four selective primers. Seventeen polymorphic fragments were identified between 500 and 1,000 bp in size compared with nine polymorphic fragments between 50 and 500 bp. Using the 500- to 1,000-bp analysis, a level of discrimination similar to that of IS6110 typing was achieved which, unlike the IS6110 typing, was able to differentiate the two M. tuberculosis strains, each of which had only a single copy of IS6110.