ABSTRACTSeveral genetic loci have been utilized to genotype isolates ofMycobacterium tuberculosis. A shortcoming of the most commonly used method, IS6110fingerprinting, is that it does not adequately discriminate between isolates having few copies of IS6110. This study was undertaken to compare pTBN12 fingerprinting of polymorphic GC-rich repetitive sequence genes and spoligotyping of the direct repeat locus as secondary typing procedures for M. tuberculosisisolates having fewer than six copies of IS6110. A total of 88 isolates (100% of the isolates with fewer than six copies of IS6110isolated in Arkansas during 1996 and 1997) were included in this study. Among the 88 isolates, 34 different IS6110patterns were observed, 10 of which were shared by more than 1 isolate, involving a total of 64 isolates. The 64 isolates were subdivided into 13 clusters (containing 37 isolates) and 27 unique isolates based on a combination of IS6110and pTBN12 fingerprinting and into 11 clusters (containing 51 isolates) and 13 unique isolates based on a combination of IS6110fingerprinting and spoligotyping. Identical spoligotypes were found among isolates having different IS6110patterns, as well as among isolates showing different pTBN12 patterns. In contrast, all isolates that had different IS6110patterns were found to be unique by pTBN12 typing. The clustering rate was 73, 58, and 42%, respectively, for IS6110fingerprinting alone, IS6110fingerprinting and spoligotyping combined, and IS6110and pTBN12 combined fingerprinting. The data indicate that the pTBN12 method has greater discriminating power among low-copy-number isolates than does spoligotyping.