Supportive Data

The correlation between pncA sequencing results and in vitro broth susceptibility test results was evaluated using 21 reference strains of Mycobacterium tuberculosis complex with known broth susceptibility profiles. Nine of 21 isolates were from the American Type Culture Collection (ATCC) and 12 of 21 isolates were from completed and closed Proficiency Testing (PT) testing events from the Center for Disease Control and Prevention (CDC), the College of American Pathologists (CAP), or the New York State Department of Health. Isolates demonstrating a polymorphism by sequencing were resequenced and all isolates had identical results between the first and second sequencing evaluation. Results are presented in Table 1.

Table 1. Accuracy of pncA Sequencing for Reference/PT Isolates

(a)SNP=single nucleotide polymorphism; see Table 3 for a description of the silent SNPs detected; a silent SNP does not result in an amino acid change.

pncA sequencing was also compared to a US Food and Drug Administration (FDA)-approved, rapid broth method(VersaTREK, TREK Diagnostic Systems) for 141 Mycobacterium tuberculosis complex isolates consisting of 96 clinical isolates and 45 reference strains (ATCC and closed PT). Any discordant results were resolved by additional testing using either the BACTEC 460 or BACTEC MGIT 960 broth methods (Becton Dickinson), which are also FDA-approved. Any isolate that had a polymorphism or that had a sequencing result that did not correlate with the broth susceptibility testing result was resequenced and identical results were found for all isolates between the first and second sequencing run. See Table 2 for pncA sequencing versus arbitrated broth susceptibility testing().

Table 2. Accuracy of pncA Sequencing vs Arbitrated Broth Susceptibility Testing

(a) for 30 isolates with discrepant VersaTREK broth and pncA sequencing results, a second broth method (either BACTEC MGIT 960 or BACTEC 460TB) was performed to determine whether the VersaTREK or sequencing result was correct.

-Sensitivity versus arbitrated broth methods=102/102 x 100=100%

-Specificity vs arbitrated broth methods=39/39 x 100=100%

-Very major error rate=0%

-Major error rate=0%

Table 3 provides a list of the pncA polymorphisms found in the validation of this method.

Table 3. pncA Nucleotide Polymorphisms Detected In House During Validation

nt=nucleotide

Silent SNPs were seen at nt positions 195, 222, 306, 408

Interday precision was evaluated by sequencing Mycobacterium tuberculosis (ATCC 27294, also known as H37Rv, PZA susceptible), Mycobacterium bovis (ATCC 19210, PZA resistant), and water (negative control) 12 times over 10 days. Mycobacterium tuberculosis ATCC 27294 gave a 100% match to the wildtype (wt) pncA sequence 12 of 12 times with good specimen quality scores (≥37) and an average consensus length of 682 + /-15 bases. Similarly, Mycobacterium bovis ATCC 19210 had a SNP present at pncA amino acid position 169, which is consistent with published literature reports for this organism. The 169 SNP was seen 12 of 12 times with good specimen quality scores (≥40) and an average consensus length of 701 ±9 bases. Interday precision was done by 2 operators using 2 ABI sequencers (Applied Biosystems) and no interoperator or interinstrument differences in performance were noted.