Validation of an Oligonucleotide Ligation Assay for Quantification of Human Immunodeficiency Virus Type 1 Drug-Resistant Mutants by Use of Massively Parallel Sequencing

Global HIV treatment programs need sensitive and affordable tests to monitor HIV drug resistance. We comparedmutant detec-tion by the oligonucleotide ligation assay (OLA), an economical and simple test, to massively parallel sequencing. Nonnucleoside reverse transcriptase inhibitor (K103N, V106M, Y181C, and G190A) and lamivudine (M184V) resistance mutations were quanti-fied in blood-derived plasma RNA and cell DNA specimens by OLA and 454 pyrosequencing. Amedian of 1,000 HIV DNA or RNA templates (range, 163 to 1,874 templates) from blood specimens collected inMozambique (n 60) and Kenya (n 51) were analyzed at 4 codons in each sample (n 441 codons assessed). Mutations were detected at 75 (17%) codons by OLA sensi-tive to 2.0%, at 71 codons (16%; P 0.78) by pyrosequencing using a cutoff value of>2.0%, and at 125 codons (28%; P< 0.0001) by pyrosequencing sensitive to 0.1%. Discrepancies between the assays included 15 codons with mutant concentrations of2%, one at 8.8 % by pyrosequencing and not detected by OLA, and one at 69 % by OLA and not detected by pyrosequencing. The latter two cases were associated with genetic polymorphisms in the regions critical for ligation of the OLA probes and pyro-sequencing primers, respectively. Overall, mutant concentrations quantified by the twomethods correlated well across the codons tested (R2> 0.8). Repeat pyrosequencing of 13 specimens showed reproducible detection of 5/24 mutations at<2 % and 6/6 at>2%. In conclusion, the OLA and pyrosequencing performed similarly in the quantification of nonnucleoside reverse transcriptase inhibitor and lamivudine mutations present at>2 % of the viral population in clinical specimens. While pyrose-quencing was more sensitive, detection of mutants below 2%was not reproducible