ABSTRACT
The ability of next-generation sequencing (NGS) technologies to detect low frequency HIV-1 drug resistance mutations (DRMs) not detected by dideoxynucleotide Sanger sequencing has potential advantages for improved patient outcomes. We compared the performance of an in vitro diagnostic (IVD) NGS assay, the Sentosa SQ HIV genotyping assay for HIV-1 genotypic resistance testing, with Sanger sequencing on 138 protease/reverse transcriptase (RT) and 39 integrase sequences. The NGS assay used a 5% threshold for reporting low-frequency variants. The level of complete plus partial nucleotide sequence concordance between Sanger sequencing and NGS was 99.9%. Among the 138 protease/RT sequences, a mean of 6.4 DRMs was identified by both Sanger and NGS, a mean of 0.5 DRM was detected by NGS alone, and a mean of 0.1 DRM was detected by Sanger sequencing alone. Among the 39 integrase sequences, a mean of 1.6 DRMs was detected by both Sanger sequencing and NGS and a mean of 0.15 DRM was detected by NGS alone. Compared with Sanger sequencing, NGS estimated higher levels of resistance to one or more antiretroviral drugs for 18.2% of protease/RT sequences and 5.1% of integrase sequences. There was little evidence for technical artifacts in the NGS sequences, but the G-to-A hypermutation was detected in three samples. In conclusion, the IVD NGS assay evaluated in this study was highly concordant with Sanger sequencing. At the 5% threshold for reporting minority variants, NGS appeared to attain a modestly increased sensitivity for detecting low-frequency DRMs without compromising sequence accuracy.
KEYWORDS: antiviral drug resistance, human immunodeficiency virus, integrase, mutations, next-generation sequencing, proteases, reverse transcriptase
INTRODUCTION
Two factors motivate the increased use of next-generation sequencing (NGS) for diagnosing HIV-1 drug resistance in both research and clinical settings. First, the cost of NGS can be considerably lower than that of dideoxynucleotide Sanger sequencing should a sufficient number of samples be tested in the same sequencing run (1, 2). Second, the ability of NGS to detect low-frequency drug resistance mutations (DRMs) not detected by Sanger sequencing has potential advantages for improved patient outcomes (3).
However, Sanger sequencing has been used for 2 decades to diagnose HIV-1 drug resistance and has been shown to be highly reproducible and interpretable in clinical settings. In contrast, NGS technologies have been rapidly evolving, resulting in changes to both laboratory and bioinformatics analysis protocols (4). Moreover, no in vitro diagnostic (IVD) assay for HIV-1 genotypic resistance testing using NGS has been approved by the U.S. Food and Drug Administration (FDA). In this study, we compared the performance of an IVD NGS assay for HIV-1 genotypic resistance testing with the current standard approach using direct PCR Sanger sequencing.
MATERIALS AND METHODS
Patients and samples.
We selected remnant cryopreserved plasma samples from individuals undergoing Sanger sequencing at the Stanford University Diagnostic Virology Laboratory for a blind comparison with the Vela Diagnostics Sentosa SQ HIV genotyping NGS assay. We selected samples that by Sanger sequencing had many common nucleoside reverse transcriptase (RT) inhibitor (NRTI)-, nonnucleoside RT inhibitor (NNRTI)-, protease (PR) inhibitor (PI) resistance-, and integrase (IN) strand transfer inhibitor (INSTI)-associated DRMs. Samples were obtained between 2001 and 2016 and were selected if their plasma HIV-1 RNA level (viral load [VL]) was ≥2.0 log copies/ml. The study was approved by the Stanford University Institutional Review Board.
Dideoxynucleoside Sanger sequencing.
Dideoxynucleoside PCR Sanger sequencing for patient management was performed separately for PR/RT and IN sequences. The PR/RT sequences encompassed the entire PR and RT codons 1 to 300. IN sequences encompassed the entire IN. Plasma samples (400 μl) were ultracentrifuged, and the pellet was subjected to RNA extraction and SuperScript III one-step RT-PCR (Thermo Fisher Scientific) followed by a second PCR. The complete set of primers for Sanger sequencing is listed in Table S1 in the supplemental material. Bidirectional sequencing was performed using BigDye Terminators (Thermo Fisher Scientific), with the products being resolved electrophoretically on an ABI 3730 sequencer. A mixture was defined as a position having a secondary peak that comprises at least 20% of the area under the curve and evidence of a mixture in both directions.
NGS.
The Sentosa SQ HIV genotyping assay is a next-generation sequencing (NGS)-based integrated work flow, comprising kits for RNA extraction, HIV-1 library preparation, and sequencing; a robotic liquid-handling system for RNA extraction and library preparation; Ion Torrent's instruments for deep sequencing; and data analysis and reporting software. The Sentosa SQ HIV genotyping NGS assay work flow is highly automated and requires about 3.5 h of hands-on time with a total turnaround time of about 27 h. The assay processes up to 15 plasma samples (730 μl per sample) simultaneously. The system sequences the entire PR sequence, the first 376 amino acids of the RT sequence, and the entire IN sequence. In this study, the median sequencing coverage for PR/RT and IN was 3,243 reads (interquartile range [IQR], 1,585 to 6,409 reads). In its current form and for the purposes of this study, the assay exports a FASTA file containing a single consensus nucleotide sequence in which positions containing nucleotide mixtures with variants present at or above 5% are represented as IUPAC ambiguities.
Data analysis.
Because many more samples underwent Sanger sequencing of the PR/RT than the IN sequence, we performed separate analyses for the PR/RT and IN sequences. Complete nucleotide concordance was defined as both Sanger sequencing and NGS identifying the same nucleotide or ambiguity code at a position. Partial nucleotide discordance was defined as one method identifying a nucleotide mixture and the other identifying one of the mixture's components. Complete nucleotide discordance was defined as both methods identifying different nonambiguous nucleotides or ambiguous nucleotides that were nonoverlapping.
For concatenated PR/RT sequences and for IN sequences, we submitted the Sanger sequencing and NGS FASTA files to the Stanford HIV Drug Resistance Database (HIVDB) genotypic resistance interpretation program and selected the “Sequence quality” and “Drug resistance” spreadsheet output options (5). We compared the rows in the “Sequence quality” spreadsheets generated from the Sanger sequencing and NGS sequences to determine concordance for detecting (i) mutations, defined as amino acid differences from the subtype B consensus amino acid sequence (https://www.hiv.lanl.gov/content/sequence/HIV/CONSENSUS/Consensus.html); (ii) DRMs, defined as mutations assigned a penalty score by the HIVDB interpretation system (5), with DRMs detected by Sanger sequencing and NGS being considered concordant if they were present regardless of what other mutations were also present; (iii) the number of signature APOBEC mutations, defined as mutations specific for APOBEC-mediated RNA editing, with three or more being associated with a high probability of G-to-A hypermutation (6); and (iv) the proportion of positions with highly unusual mutations, defined as having a prevalence of <0.01% in HIVDB (6) and not being a known DRM or signature APOBEC mutation.
We compared the rows in the “Drug resistance summary” spreadsheets generated from the Sanger sequencing and NGS sequences to determine the concordance for the detection of categorical drug resistance interpretations for the most commonly used antiretrovirals (ARVs): the NRTIs lamivudine (3TC) and emtricitabine (FTC) (which were treated as one drug), abacavir (ABC), zidovudine (AZT), and tenofovir (TDF); the NNRTIs efavirenz (EFV), etravirine (ETR), and rilpivirine (RPV); the PIs atazanavir (ATV), darunavir (DRV), and lopinavir (LPV); and the INSTIs dolutegravir (DTG), elvitegravir (EVG), and raltegravir (RAL). There were five predicted drug resistance interpretation levels: susceptible, potential low-level resistance, low-level resistance, intermediate resistance, and high-level resistance (5). These levels were used to compare the drug resistance interpretations reported by NGS and Sanger sequencing.
Accession number(s).
The complete set of 138 PR/RT and 39 IN sequences determined by NGS and Sanger sequencing, annotated by sample identifier and sequencing method, has been submitted to GenBank and may be found under accession numbers KY190203, KY190186, KY190168, KY190161, and MG800345 to MG800626 (Data Set S3).
RESULTS
Patients and samples.
Overall, 143 samples were sequenced by both NGS and Sanger sequencing, including 104 that underwent PR/RT sequencing, 34 that underwent PR/RT and IN sequencing, and 5 that underwent just IN sequencing. Thus, 138 samples underwent PR/RT sequencing and 34 underwent IN sequencing by both Sanger sequencing and NGS. In all samples, both Sanger sequencing and NGS encompassed the entire PR sequence, the first 250 positions of the RT sequence, and the entire IN sequence. Twelve samples (8.4% of 143) had non-B subtypes, including subtypes C (n = 4), CRF01_AE (n = 3), A (n = 2), G (n = 1), CRF02_AG (n = 1), and CRF07_BC (n = 1). The VL was available for 127 (88.8%) of 143 samples. The median VL for these samples was 3.8 log10 copies/ml (IQR, 3.1 to 4.5 log10 copies/ml).
Treatment histories were available for 112 individuals undergoing PR/RT sequencing. Among these individuals, the median numbers of NRTIs, NNRTIs, and PIs were 4, 1, and 2, respectively. Among 29 individuals undergoing IN sequencing for whom the treatment history was available, the median number of INSTIs received was 1.
Table 1 lists each of the NRTI, NNRTI, PI, and INSTI DRMs detected by Sanger sequencing according to its frequency in the tested samples. The samples contained 137 distinct DRMs: 40 NRTI-, 34 NNRTI-, 41 PI-, and 22 INSTI-associated DRMs. Table S2 in the supplemental material lists the DRMs and their penalty scores for the most commonly used ARVs. There was a high correlation between the proportions of DRMs in this data set and those from ARV-treated persons in HIVDB (r2 > 0.8; P < 0.001).
TABLE 1.
NRTI, NNRTI, PI, and INSTI DRMs detected by Sanger sequencing genotypic resistance testing in 138 PR/RT and 39 IN sequencesa
| Inhibitor and DRM | No. | % | % HIVDB |
|---|---|---|---|
| NRTI | |||
| M184V | 67 | 48.6 | 53.6 |
| M41L | 52 | 37.7 | 27.3 |
| D67N | 41 | 29.7 | 26 |
| T215Y | 38 | 27.5 | 24.6 |
| L210W | 31 | 22.5 | 16.6 |
| K70R | 29 | 21 | 17.6 |
| K219Q | 27 | 19.6 | 10.5 |
| L74V | 20 | 14.5 | 7.5 |
| K65R | 19 | 13.8 | 6 |
| T215F | 19 | 13.8 | 9.1 |
| L74I | 10 | 7.2 | 3.7 |
| K219E | 10 | 7.2 | 5.8 |
| E44D | 10 | 7.2 | 7.2 |
| T69D | 9 | 6.5 | 5.4 |
| D67G | 8 | 5.8 | 2.3 |
| T69ins | 7 | 5.1 | 0.7 |
| A62V | 7 | 5.1 | 4.5 |
| T215D | 5 | 3.6 | 0.5 |
| K219R | 4 | 2.9 | 2.4 |
| M184I | 4 | 2.9 | 2 |
| K219N | 4 | 2.9 | 2.6 |
| Y115F | 3 | 2.2 | 3 |
| V75M | 3 | 2.2 | 3.2 |
| E44A | 3 | 2.2 | 1.3 |
| K70G | 3 | 2.2 | 0.3 |
| K70E | 3 | 2.2 | 1 |
| T215E | 3 | 2.2 | 0.1 |
| K70Q | 2 | 1.4 | 0.3 |
| E40F | 2 | 1.4 | 0.5 |
| V75I | 2 | 1.4 | 2.8 |
| F77L | 2 | 1.4 | 1.7 |
| D67H | 2 | 1.4 | 0.2 |
| T215L | 1 | 0.7 | 0.1 |
| D67E | 1 | 0.7 | 0.4 |
| T215I | 1 | 0.7 | 1.3 |
| Q151M | 1 | 0.7 | 2.6 |
| F116Y | 1 | 0.7 | 2.1 |
| T215A | 1 | 0.7 | 0.1 |
| V75T | 1 | 0.7 | 1.1 |
| T215C | 1 | 0.7 | 0.5 |
| NNRTI | |||
| K103N | 46 | 33.3 | 29.5 |
| Y181C | 28 | 20.3 | 16.9 |
| K101E | 17 | 12.3 | 6.9 |
| G190A | 14 | 10.1 | 13.5 |
| V108I | 14 | 10.1 | 7.7 |
| P225H | 12 | 8.7 | 4.2 |
| H221Y | 11 | 8 | 6.7 |
| L100I | 9 | 6.5 | 3.4 |
| A98G | 8 | 5.8 | 6.4 |
| K101P | 7 | 5.1 | 1.1 |
| K238T | 6 | 4.3 | 1.9 |
| Y188L | 6 | 4.3 | 3.7 |
| Y181I | 4 | 2.9 | 0.7 |
| E138A | 4 | 2.9 | 3.3 |
| G190S | 4 | 2.9 | 2 |
| Y188F | 3 | 2.2 | 0.1 |
| V179F | 3 | 2.2 | 0.2 |
| K101H | 2 | 1.4 | 0.9 |
| Y188C | 2 | 1.4 | 0.6 |
| E138K | 2 | 1.4 | 0.5 |
| G190E | 2 | 1.4 | 0.4 |
| K238N | 2 | 1.4 | 0.3 |
| F227L | 2 | 1.4 | 2.7 |
| K103S | 2 | 1.4 | 1.1 |
| M230L | 2 | 1.4 | 1.7 |
| Y181F | 1 | 0.7 | 0.1 |
| Y188H | 1 | 0.7 | 0.3 |
| V179L | 1 | 0.7 | 0.1 |
| E138Q | 1 | 0.7 | 1.1 |
| V106A | 1 | 0.7 | 1.2 |
| V179E | 1 | 0.7 | 1.1 |
| V106M | 1 | 0.7 | 5.1 |
| L100V | 1 | 0.7 | 0.1 |
| V179D | 1 | 0.7 | 2.4 |
| PI | |||
| L90M | 27 | 19.6 | 25.8 |
| M46I | 19 | 13.8 | 19.7 |
| V82A | 17 | 12.3 | 20.5 |
| L10F | 14 | 10.1 | 7.6 |
| I54V | 14 | 10.1 | 21.8 |
| L33F | 11 | 8 | 11.5 |
| K43T | 10 | 7.2 | 0.1 |
| I84V | 9 | 6.5 | 1 |
| N88D | 9 | 6.5 | 4.9 |
| D30N | 8 | 5.8 | 5.3 |
| V32I | 6 | 4.3 | 4.6 |
| I47V | 6 | 4.3 | 0.1 |
| G73S | 5 | 3.6 | 6.6 |
| L23I | 5 | 3.6 | 1.3 |
| M46L | 5 | 3.6 | 8 |
| L89V | 4 | 2.9 | 3.2 |
| V82T | 4 | 2.9 | 2.1 |
| G48V | 3 | 2.2 | 2.8 |
| V11I | 3 | 2.2 | 2.7 |
| F53L | 3 | 2.2 | 5.6 |
| I54M | 3 | 2.2 | 1.9 |
| I50V | 3 | 2.2 | 1.3 |
| I54L | 3 | 2.2 | 2.6 |
| G73T | 3 | 2.2 | 1.9 |
| K20T | 2 | 1.4 | 0.1 |
| G73A | 2 | 1.4 | 0.4 |
| L76V | 2 | 1.4 | 3.5 |
| Q58E | 2 | 1.4 | 6.3 |
| N88S | 2 | 1.4 | 1.5 |
| V82S | 2 | 1.4 | 0.9 |
| T74P | 2 | 1.4 | 1.8 |
| I54S | 1 | 0.7 | 0.5 |
| L24I | 1 | 0.7 | 5.2 |
| I50L | 1 | 0.7 | 0.1 |
| I54T | 1 | 0.7 | 0.6 |
| G48M | 1 | 0.7 | 0.4 |
| I47A | 1 | 0.7 | 0.5 |
| V82M | 1 | 0.7 | 0.4 |
| I54A | 1 | 0.7 | 1 |
| L24M | 1 | 0.7 | 0.1 |
| V82F | 1 | 0.7 | 1.6 |
| INSTI | |||
| N155H | 11 | 28.2 | 29.9 |
| G140S | 5 | 12.8 | 24.6 |
| T97A | 5 | 12.8 | 12.2 |
| E92Q | 5 | 12.8 | 6 |
| Q148R | 5 | 12.8 | 7.1 |
| Q148H | 4 | 10.3 | 22 |
| E138K | 4 | 10.3 | 0.5 |
| G163K | 3 | 7.7 | 1.7 |
| G163R | 2 | 5.1 | 7 |
| G140A | 2 | 5.1 | 1.7 |
| E138A | 2 | 5.1 | 3.3 |
| E157Q | 2 | 5.1 | 7.4 |
| Y143R | 2 | 5.1 | 6.6 |
| Y143C | 2 | 5.1 | 3.9 |
| S147G | 2 | 5.1 | 1.5 |
| T66I | 1 | 2.6 | 1.2 |
| L74M | 1 | 2.6 | 0 |
| L74I | 1 | 2.6 | 3.7 |
| Q95K | 1 | 2.6 | 1.4 |
| S230R | 1 | 2.6 | 3.5 |
| T66A | 1 | 2.6 | 0.5 |
| Q148N | 1 | 2.6 | 0.2 |
No., number of DRMs among the clinical samples; %, percentage of the samples containing the DRM; % HIVDB, prevalence of the DRM among ARV-treated individuals in the Stanford HIV Drug Resistance Database.
Nucleotide sequence concordance and sequence ambiguities.
In both the PR/RT and IN sequences, the nucleotide sequences determined by Sanger sequencing and NGS were highly concordant: 98.37% of nucleotides were identical, 1.58% were partially discordant, and 0.05% were completely discordant. Not surprisingly, the median number of IUPAC ambiguities was significantly higher in the sequences determined by NGS than those determined by Sanger sequencing: 2.1 (IQR, 1.0 to 3.5) versus 1.1 (IQR, 0.4 to 1.8) (P < 0.001, Mann-Whitney U test). The median 1.1% level of IUPAC ambiguities detected by Sanger sequencing is similar to the proportions of 0.7%, 1.1%, and 1.1% reported from three laboratories performing Sanger sequencing on similar heavily treated patients (7, 8). The higher proportion of ambiguities by NGS results from its lower threshold for detecting variants.
Mutations, DRMs, and drug resistance interpretations.
Among the 138 PR/RT sequences, a mean of 24.2 mutations, defined as differences from the subtype B amino acid consensus sequence, was detected by both Sanger sequencing and NGS (Table 2). NGS identified a mean of 4.4 mutations not detected by Sanger sequencing, whereas Sanger sequencing identified a mean of 0.6 mutation not detected by NGS (P < 0.001, Student's t test). Among the 39 IN sequences, a mean of 12.1 mutations was identified by both Sanger sequencing and NGS (Table 2). NGS identified a mean of 2.7 mutations not detected by Sanger sequencing, whereas Sanger sequencing identified a mean of 1.2 mutations not detected by NGS (P = 0.03, Student's t test).
TABLE 2.
Mean number of amino acid mutations and DRMs detected by standard Sanger sequencing genotypic resistance testing and NGSa
| Gene and mutation categoryb | Mean value by: |
P valuec | |
|---|---|---|---|
| Sanger sequencing | NGS | ||
| PR/RT (n = 138 sequences) | |||
| No. of mutations | |||
| Shared | 24.20 | 24.20 | NA |
| Unique | 0.64 | 4.38 | <0.001 |
| No. of DRMs | |||
| Shared | 6.39 | 6.39 | NA |
| Unique | 0.10 | 0.54 | <0.001 |
| IN (n = 39 sequences) | |||
| No. of mutations | |||
| Shared | 12.05 | 12.05 | NA |
| Unique | 1.18 | 2.69 | 0.03 |
| No. of DRMs | |||
| Shared | 1.59 | 1.59 | NA |
| Unique | 0.03 | 0.15 | 0.09 |
Amino acid mutations are defined as differences from the subtype B consensus reference sequence. Drug resistance mutations (DRMs) are mutations that receive a penalty score in the HIVDB drug resistance interpretation program.
Shared mutations and DRMs are those that were detected by both NGS and Sanger sequencing. Unique mutations and DRMs are those that were detected by either NGS or Sanger sequencing but not by both.
P values were determined by Student's t test. NA, not applicable.
Among the 138 PR/RT sequences, a mean of 6.4 drug resistance mutations (DRMs) was identified by both Sanger sequencing and NGS (Table 2). NGS identified a mean of 0.5 mutation not detected by Sanger sequencing, whereas Sanger sequencing detected a mean of 0.1 mutation not detected by NGS (P < 0.001, Student's t test). Among the 39 IN sequences, a mean of 1.6 DRMs was identified by both Sanger sequencing and NGS (Table 2). NGS identified a mean of 0.2 DRM detected by Sanger sequencing, whereas Sanger sequencing identified a mean of 0.03 DRM not detected by NGS (P = 0.09, Student's t test). Among the 11 DRMs detected by Sanger sequencing but not NGS, each was present as a mixture by Sanger sequencing and 6 were present below the 5% threshold when the raw NGS reads were examined.
On average, there was no significant difference between Sanger sequencing and NGS in the predicted levels of resistance to any of the NRTIs, NNRTIs, PIs, and INSTIs (Table 3). However, for the 138 PR/RT samples, NGS detected a higher level of NRTI resistance for 6.2% (n = 9) of samples, a higher level of NNRTI resistance for 11.2% (n = 16) of samples, and a higher level of PI resistance for 4.1% (n = 6) of samples than Sanger sequencing (Tables 4 to 6). In contrast, Sanger sequencing detected higher levels of NRTI, NNRTI, and PI resistance for 0.7% (n = 1), 1.4% (n = 2), and 0.7% (n = 1) of samples, respectively. Overall, 26 (18.2%) individuals with PR/RT sequences had a higher level of predicted resistance to one or more RTIs or PIs based on NGS. Four (2.8%) had a higher level of predicted resistance based on Sanger sequencing. For the 39 IN samples, NGS detected a higher level of INSTI resistance for 5.1% (n = 2) of the samples (Table 7).
TABLE 3.
Mean levels of predicted reduced ARV susceptibility of the 138 PR/RT and 39 IN samples according to genotypic resistance testing by Sanger sequencing versus NGS
| Drug class | Druga | Mean susceptibility levelb ± SD |
P valuec | |
|---|---|---|---|---|
| Sanger sequencing | NGS | |||
| NRTIs (n = 138 sequences) | 3FTC | 3.43 ± 1.79 | 3.46 ± 1.79 | 0.9 |
| ABC | 3.80 ± 1.57 | 3.78 ± 1.60 | 0.9 | |
| AZT | 3.16 ± 1.86 | 3.13 ± 1.83 | 0.9 | |
| TDF | 3.12 ± 1.58 | 3.12 ± 1.61 | 1.0 | |
| NNRTIs (n = 138 sequences) | EFV | 3.33 ± 1.86 | 3.41 ± 1.86 | 0.7 |
| ETR | 2.54 ± 1.61 | 2.64 ± 1.64 | 0.6 | |
| RPV | 2.94 ± 1.85 | 3.06 ± 1.88 | 0.3 | |
| PIs (n = 138 sequences) | ATV/r | 2.23 ± 1.69 | 2.26 ± 1.70 | 0.9 |
| DRV/r | 1.43 ± 1.03 | 1.43 ± 1.05 | 1.0 | |
| LPV/r | 2.09 ± 1.58 | 2.12 ± 1.59 | 0.9 | |
| INSTIs (n = 39 sequences) | DTG | 2.31 ± 1.25 | 2.33 ± 1.27 | 0.9 |
| EVG | 3.77 ± 1.70 | 3.79 ± 1.71 | 0.9 | |
| RAL | 3.79 ± 1.67 | 3.79 ± 1.67 | 1.0 | |
3FTC, lamivudine and emtricitabine; ABC, abacavir; AZT, zidovudine; TDF, tenofovir; EFV, efavirenz; ETR, etravirine; RPV, rilpivirine; ATV/r, ritonavir-boosted atazanavir; DRV/r, ritonavir-boosted darunavir; LPV/r, ritonavir-boosted lopinavir; DTG, dolutegravir; EVG, elvitegravir; RAL, raltegravir.
Susceptibility levels were given the following scores: 1 for susceptible, 2 for potential low-level resistance, 3 for low-level resistance, 4 for intermediate resistance, and 5 for high-level resistance.
P values were determined by Student's t test.
TABLE 4.
NRTI resistance mutations and estimated NRTI resistance levels for sequences with differences in genotypic resistance testing results by Sanger sequencing and NGS
| Sample group and sample identifier | DRM detected by: |
Sanger sequencing score → NGS scorea |
|||||
|---|---|---|---|---|---|---|---|
| Sanger sequencing and NGS | Sanger sequencing alone | NGS alone | 3FTC | ABC | AZT | TDF | |
| Samples for which NGS reported higher levels of resistance to ≥1 NRTIs | |||||||
| B3-31 | 41L, 215Y | 184V | 1 → 5 | 4 | 5 → 4 | 3 | |
| B3-10 | 184V | 65R, 74V | 5 | 3 → 5 | 1 | 1 → 4 | |
| B3-28 | 41L, 210W, 215Y | 67N | 3 | 5 | 5 | 4 → 5 | |
| B1-10 | 41L, 44D, 184V, 210W, 215Y | 67N | 5 | 5 | 5 | 4 → 5 | |
| B2-05 | 40F, 41L, 74V, 210W, 215Y, 219Q | 67N | 69ins | 3 → 4 | 5 | 5 | 5 |
| B3-25 | 65R, 184V | 70E | 5 | 5 | 1 | 4 → 5 | |
| B3-02 | 41L, 67N, 70R, 184V, 219E | 215I | 5 | 5 | 5 | 3 → 4 | |
| B2-17 | 184V | 41L, 62V | 5 | 3 | 1 → 2 | 1 | |
| B1-51 | 65R, 67N, 184V, 219E | 70R, 215I | 5 | 5 | 1 → 5 | 5 | |
| Samples for which Sanger sequencing reported higher levels of resistance to ≥1 NRTIs, B2-15 | 67N, 215I, 219E | 70R | 2 → 1 | 4 → 3 | 5 → 4 | 4 → 3 | |
| Samples for which Sanger sequencing or NGS detected different DRMs but for which levels of NRTI resistance were the same | |||||||
| B1-42 | 41L, 62V, 184V, 210W, 215Y | 219R | 5 | 5 | 5 | 4 | |
| B2-31 | 67N, 219Q | 215L | 215V | 1 | 3 | 4 | 3 |
| B2-01 | 67N, 70R, 184V, 215Y, 219Q | 74I | 5 | 5 | 5 | 4 | |
| B2-38 | 41L, 69ins, 210W, 215Y | 62V | 4 | 5 | 5 | 5 | |
| B2-16 | 40F, 41L, 67N, 75 M, 210W, 215Y | 44A, 75I, 219R | 3 | 5 | 5 | 5 | |
| B2-04 | 67N, 70R, 219Q | 215F | 215V | 2 | 4 | 5 | 4 |
| B1-44 | 65R, 74I, 184V, 219R | 70Q | 5 | 5 | 1 | 5 | |
| B1-18 | 74V, 184V | 219R | 5 | 5 | 1 | 1 | |
Predicted levels of drug resistance according to the HIVDB genotypic resistance interpretation system, which were scored as follows: 1 for susceptible, 2 for potential low-level resistance, 3 for low-level resistance, 4 for intermediate resistance, and 5 for high-level resistance. For an additional 5 sequences, NGS detected two mutations at a DRM position for which Sanger sequencing detected a single mutation. However, the additional mutation did not result in a change in the estimated level of drug resistance. 3FTC, lamivudine and emtricitabine; ABC, abacavir; AZT, zidovudine; TDF, tenofovir.
TABLE 5.
NNRTI resistance mutations and estimated NNRTI resistance levels for sequences with differences in genotypic resistance testing results by Sanger sequencing and NGS
| Sample group and sample identifier | DRM detected by: |
Sanger sequencing score → NGS scorea |
||||
|---|---|---|---|---|---|---|
| Sanger sequencing and NGS | Sanger sequencing alone | NGS alone | EFV | ETR | RPV | |
| Samples for which NGS reported higher levels of resistance to ≥1 NNRTIs | ||||||
| B2-16 | 103N | 1 → 5 | 1 | 1 | ||
| B2-35 | 138A | 103N | 1 → 5 | 2 | 3 | |
| B3-33 | 103N | 181C | 5 | 1 → 4 | 1 → 4 | |
| B1-49 | 101E | 190S | 3 → 5 | 3 → 4 | 4 → 5 | |
| B3-31 | 181I | 103N | 4 → 5 | 5 | 5 | |
| B2-07 | 101H, 188L, 238T | 98G, 179D, 181C, 190A | 5 | 3 → 5 | 5 | |
| B1-51 | 181C | 98G, 101E | 4 → 5 | 4 → 5 | 4 → 5 | |
| B3-26 | 103N, 181C | 190A, 238T | 5 | 4 | 4 → 5 | |
| B1-43 | 190S | 101E | 5 | 2 → 4 | 3 → 5 | |
| B2-11b | 190E | 230I | 5 | 4 → 5 | 5 | |
| B1-35 | 103N, 181C | 221Y | 5 | 4 | 4 → 5 | |
| B3-34 | 103N, 179L | 181C | 5 | 2 → 4 | 3 → 5 | |
| B1-33 | 103N, 238N | V179L | 5 | 1 → 2 | 1 → 3 | |
| B2-01 | 181C, 188H, 190A | V108I, H221Y | 5 | 4 → 5 | 5 | |
| B1-44 | 103N, 108I, 225H | 221Y | 5 | 1 → 2 | 1 → 3 | |
| B1-16 | 103N, 181C, 188C | 106A, 190A | 5 | 4 | 4 → 5 | |
| Samples for which Sanger sequencing reported higher levels of resistance to ≥1 NNRTIs | ||||||
| B3-29 | 101E, 190A | 103N, 181C | 5 | 5 → 4 | 5 | |
| B1-37 | 181C | 101E, 103N | 221Y | 5 → 4 | 4 | 5 |
| Samples for which Sanger sequencing or NGS detected different DRMs but for which levels of NNRTI resistance were the same | ||||||
| B2-32 | 100I, 238T | 103N | 5 | 4 | 5 | |
| B3-42 | 188L | 103N | 5 | 2 | 5 | |
| B3-36 | 98G, 101E, 103N, 190A | 225H | 5 | 4 | 5 | |
| B2-28 | 179E, 188L | 108I | 5 | 3 | 5 | |
| B2-29 | 98G, 103N, 108I, 221Y, 227L | 238N | 5 | 3 | 4 | |
| B1-04 | 100I, 103N, 225H | 108I | 5 | 4 | 5 | |
| B2-49 | 101P, 230L, 238N | 101E, 138K, 181C | 5 | 5 | 5 | |
Predicted levels of drug resistance according to the HIVDB genotypic resistance interpretation system, which were scored as follows: 1 for susceptible, 2 for potential low-level resistance, 3 for low-level resistance, 4 for intermediate resistance, and 5 for high-level resistance. EFV, efavirenz; ETR, etravirine; RPV, rilpivirine.
Sample B2-11 had evidence for an APOBEC-mediated G-to-A hypermutation. Both G190E (which was also detected by Sanger sequencing) and M230I (which was detected only by NGS) are DRMs that occur in the APOBEC dinucleotide context.
TABLE 6.
PI resistance mutations and estimated PI resistance levels for sequences with differences in genotypic resistance testing results by Sanger sequencing and NGS
| Sample group and sample identifier | DRM detected by: |
Sanger sequencing score → NGS scorea |
||||
|---|---|---|---|---|---|---|
| Sanger sequencing and NGS | Sanger sequencing alone | NGS alone | ATV/r | DRV/r | LPV/r | |
| Samples for which NGS reported higher levels of resistance to ≥1 PIs | ||||||
| B3-31 | 10F, 82A | 46L | 3 → 4 | 1 | 4 | |
| B1-10 | 46L, 82A, 84V, 90M | 46I, 54M | 5 | 3 → 4 | 5 | |
| B1-42 | 46I, 90M | 10F | 4 | 1 | 3 → 4 | |
| B2-02 | 33F, 50V, 89V | 46I | 1 → 3 | 4 | 4 | |
| B3-28 | 90M | 1 → 3 | 1 | 1 → 3 | ||
| B2-04 | 88S | 10F, 46I | 5 | 1 | 1 → 3 | |
| Samples for which Sanger sequencing reported higher levels of resistance to ≥1 PIs, B3-19 | 46L | 2 → 1 | 1 | 2 → 1 | ||
| Samples for which Sanger sequencing or NGS detected different DRMs but for which levels of PI resistance were the same | ||||||
| B1-07 | 10F, 11I, 32I, 46I, 54 M, 73A, 84V, 90M | 47V | 5 | 5 | 5 | |
| B2-32 | 48V, 54S, 82A | 53L | 5 | 1 | 5 | |
Predicted levels of drug resistance according to the HIVDB genotypic resistance interpretation system, which were scored as follows: 1 for susceptible, 2 for potential low-level resistance, 3 for low-level resistance, 4 for intermediate resistance, and 5 for high-level resistance. For an additional two sequences, NGS detected two mutations at a DRM position for which Sanger detected a single mutation. However, the additional mutation did not result in a change in the estimated level of drug resistance. ATV/r, ritonavir-boosted atazanavir; DRV/r, ritonavir-boosted darunavir; LPV/r, ritonavir-boosted lopinavir.
TABLE 7.
INSTI resistance mutations and estimated INSTI resistance levels for sequences with differences in genotypic resistance testing results by Sanger sequencing and NGS
| Sample group and sample identifier | DRM detected by: |
Sanger sequencing score → NGS scorea |
||||
|---|---|---|---|---|---|---|
| Sanger sequencing and NGS | Sanger sequencing alone | NGS alone | DTG | EVG | RAL | |
| Samples for which NGS reported higher levels of resistance to ≥1 INSTIs | ||||||
| B3-17 | 148H | 140S | 3 → 4 | 5 | 5 | |
| B3-41 | 97A, 138A, 143C | 155H | 3 | 4 → 5 | 5 | |
| Samples for which Sanger sequencing or NGS detected different DRMs but for which levels of INSTI resistance were the same | ||||||
| B1-33b | 138A, 140A, 148R | 163R | 5 | 5 | 5 | |
| B3-07 | 155H | 157Q | 2 | 5 | 5 | |
| B3-09 | 155H, 163K | 97A | 2 | 5 | 5 | |
| B3-25b | 155H | 97A, 163K | 2 | 5 | 5 | |
Predicted levels of drug resistance according to the HIVDB genotypic resistance interpretation system, which were scored as follows: 1 for susceptible, 2 for potential low-level resistance, 3 for low-level resistance, 4 for intermediate resistance, and 5 for high-level resistance. DTG, dolutegravir; EVG, elvitegravir; RAL, raltegravir.
Samples B1-33 and B3-25 had weak evidence for an APOBEC-mediated G-to-A hypermutation. G163R and G163K are accessory INSTI resistance mutations that occur in the APOBEC dinucleotide context.
Quality control comparison.
There was no significant difference in the mean number of signature APOBEC mutations per sequence between Sanger sequencing and NGS in either PR/RT sequences (0.02 for Sanger sequencing versus 0.14 for NGS; P = 0.2, paired Student's t test) or IN sequences (0.05 for Sanger sequencing versus 0.41 for NGS; P = 0.06). One PR/RT sequence and three IN sequences determined by NGS had three or more signature APOBEC mutations. Sample B2-11 had 12 signature APOBEC mutations, including six stop codons. Samples B3-25, B1-33, and B3-02 had five, four, and three signature APOBEC mutations, respectively. In contrast, no sequence determined by Sanger sequencing had more than two signature APOBEC mutations.
The PR/RT sample B2-11, which had 12 signature APOBEC mutations, including six stop codons at the default 5% threshold, had no signature APOBEC mutations or stop codons at a 20% threshold. This sample had two NNRTI resistance mutations, G190E and M230I, which are among the few NNRTI resistance mutations that can be caused by APOBEC (Table 5). G190E was detected by both Sanger sequencing and NGS, but M230I was detected only by NGS (at the 5% threshold but not the 20% threshold). This sample had a VL of 3.8 log copies/ml and a subtype B virus.
The mean number of highly unusual mutations per sequence in PR/RT was higher for the sequences determined by NGS than for the sequences determined by Sanger sequencing (0.75 versus 0.29; P = 0.002, paired Student's t test). There was no significant difference in the mean number of highly unusual mutations in the IN sequences (0.44 versus 0.38; P = 0.8, paired Student's t test). Among the 138 PR/RT sequences, the highest number of highly unusual mutations was four for Sanger sequencing. For NGS, five samples had a higher number of highly unusual mutations, including one with six and four with five highly unusual mutations.
DISCUSSION
This study is one of the largest comparisons of HIV-1 genotypic resistance testing conducted using Sanger sequencing and NGS and one of the few comparing Sanger sequencing to an IVD NGS assay. Despite the use of different methods for sample preparation and amplification, the level of complete plus partial nucleotide sequence concordance between Sanger sequencing and NGS was extraordinarily high, matching the 99.9% concordance rate reported in a previous study comparing different Sanger sequencing protocols (8). Compared with Sanger sequencing, NGS detected significantly more DRMs per sequence and estimated higher levels of resistance to one or more ARVs for 18.2% of PR/RT sequences and 5.1% of IN sequences.
Sanger sequencing usually detects HIV-1 variants present in proportions above 20%, with a range of 10% to 30%, depending on the nucleotide context (7, 9–12). In contrast, the proportion at which variants can be detected by NGS depends on the selected threshold. Although detection thresholds as low as 1% have often been used in research studies, many of these studies excluded samples with low plasma virus levels (13–17). Two recent descriptions of HIV-1 NGS analysis pipelines suggested using a detection threshold of 5% to minimize technical artifacts (18, 19).
Because of its extraordinary sensitivity, NGS is prone to two types of artifacts that occur uncommonly during Sanger sequencing: PCR error and APOBEC-mediated G-to-A hypermutation. PCR errors commonly occur during nested PCR but are rarely present at levels high enough to be detected by Sanger sequencing (20–23). In contrast, PCR errors are detected by NGS if the detection threshold is set too low. Indeed, if the number of amplifiable cDNA templates is low due to a low VL and/or inefficient RNA extraction, reverse transcription, or PCR, then much of the observed variability in an NGS sequence will represent PCR error (21, 22, 24). Indeed, studies using the Primer ID approach, which adds a unique label to each cDNA molecule, shows that the depth of NGS does not necessarily lead to the reliable detection of low-frequency variants (21, 25, 26).
NGS is also more likely than Sanger sequencing to detect low-level APOBEC-mediated G-to-A hypermutation, particularly in samples contaminated with proviral DNA (27–29). Hypermutated viruses are unlikely to be functional and have not been shown to contribute to virological failure (27–30). There are 13 DRMs that can be caused by APOBEC-mediated hypermutation, including D30N, M46I, and G73S in PR; D67N, E138K, M184I, G190SE, and M230I in RT; and E138K, G118R, and G163KR in IN. Determining whether a sequence contains an APOBEC-mediated G-to-A hypermutation is necessary for interpreting the significance of these DRMs.
We have hypothesized that an inappropriately low detection threshold for a sample can be determined in a post hoc analysis by counting its number of highly unusual and signature APOBEC mutations at different thresholds (6, 31). The presence of a large number of highly unusual mutations above a threshold suggests that the threshold is too low and that some of the low-frequency mutations are PCR errors. Likewise, the presence of a large number of signature APOBEC mutations above a threshold suggests that those DRMs that can be caused by APOBEC-mediated hypermutation may not reflect ARV selection pressure.
In this study, we identified one PR/RT sample and three IN samples with evidence for an APOBEC-mediated G-to-A hypermutation and five PR/RT samples with a slightly greater number of highly unusual mutations by NGS than by Sanger sequencing. The one hypermutated PR/RT sample was readily identifiable, and our post hoc analysis showed that a cutoff of 20% was appropriate for this sample. The relatively infrequent detection of signature APOBEC mutations and highly unusual mutations by NGS in this study is consistent with the conservative 5% threshold.
Low-frequency DRMs include those that were once dominant in an individual's virus population as a result of transmitted or acquired resistance but that, in the absence of drug pressure, were outgrown by more fit variants (32–34) and those that have emerged under selective drug pressure but that have not reached levels detectable by Sanger sequencing (27, 35). In this study of samples obtained primarily from heavily treated individuals with multiple DRMs detectable by Sanger sequencing, NGS detected 4.4 additional PR/RT mutations per sample, of which 0.5 were DRMs, and 2.7 additional IN mutations per sample, of which 0.2 were DRMs. The number of additional DRMs detected by NGS than by Sanger sequencing was higher in several other NGS studies of treated individuals that used detection thresholds lower than the threshold used in this study (13, 16, 36, 37).
The strongest evidence for the clinical significance of low-frequency DRMs has been for the CXCR5 inhibitor (38) and NNRTI (39, 40) classes. One meta-analysis and two subsequent case-control and cohort studies have shown that low-frequency NNRTI DRMs are associated with an increased risk of virological failure on a first-line NNRTI-containing regimen (15, 39, 41). The cohort study showed an increased risk of virological failure associated with low-frequency variants present at levels above 5% but not at lower levels (41). Low-frequency accessory PI-associated DRMs have not been shown to reduce the response to a first-line ritonavir-boosted ATV- or ritonavir-boosted DRV-containing regimen (42).
Low-frequency DRMs have been associated with an increased risk of virological failure in one retrospective salvage therapy study (13) but demonstrated only a trend toward increased virological failure in a second study of salvage therapy with the potent combination of ritonavir-boosted darunavir, raltegravir, and etravirine (37). Low-frequency DRMs also appear to be more likely to be clinically significant in patients with a treatment history compatible with the DRM, possibly because in this setting the low-frequency DRMs are less likely to be artifactual and more likely to be linked to compensatory mutations (16, 36, 40).
In conclusion, this study demonstrates a high concordance rate between NGS and Sanger sequencing. Despite using a conservative 5% threshold for low-frequency variants, NGS still detected a mean of an additional 0.5 PR/RT DRM and 0.1 IN DRM per sequence, resulting in a predicted increased level of resistance to one or more ARVs for 18% of PR/RT sequences and 5% of IN sequences. Additional studies are needed to elucidate the clinical significance of different low-frequency DRMs, although it is likely that those DRMs causing the greatest reductions in susceptibility to ARVs with low genetic barriers to resistance will be the most clinically relevant.
Supplementary Material
ACKNOWLEDGMENTS
P.L.T., V.V., and R.W.S. were funded in part by NIH grant AI068581 and in part by a grant from Vela Diagnostics.
P.A., C.L., E.R., C.V., M.Y., K.T., and G.M. are employees of Vela Diagnostics.
Footnotes
Supplemental material for this article may be found at https://doi.org/10.1128/JCM.00105-18.
REFERENCES
- 1.Van Laethem K, Theys K, Vandamme AM. 2015. HIV-1 genotypic drug resistance testing: digging deep, reaching wide? Curr Opin Virol 14:16–23. doi: 10.1016/j.coviro.2015.06.001. [DOI] [PubMed] [Google Scholar]
- 2.Lapointe HR, Dong W, Lee GQ, Bangsberg DR, Martin JN, Mocello AR, Boum Y, Karakas A, Kirkby D, Poon AF, Harrigan PR, Brumme CJ. 2015. HIV drug resistance testing by high-multiplex “wide” sequencing on the MiSeq instrument. Antimicrob Agents Chemother 59:6824–6833. doi: 10.1128/AAC.01490-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Stella-Ascariz N, Arribas JR, Paredes R, Li JZ. 2017. The role of HIV-1 drug-resistant minority variants in treatment failure. J Infect Dis 216:S847–S850. doi: 10.1093/infdis/jix430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Noguera-Julian M, Edgil D, Harrigan PR, Sandstrom P, Godfrey C, Paredes R. 2017. Next-generation human immunodeficiency virus sequencing for patient management and drug resistance surveillance. J Infect Dis 216(Suppl 9):S829–S833. doi: 10.1093/infdis/jix397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Paredes R, Tzou PL, van Zyl G, Barrow G, Camacho R, Carmona S, Grant PM, Gupta RK, Hamers RL, Harrigan PR, Jordan MR, Kantor R, Katzenstein DA, Kuritzkes DR, Maldarelli F, Otelea D, Wallis CL, Schapiro JM, Shafer RW. 2017. Collaborative update of a rule-based expert system for HIV-1 genotypic resistance test interpretation. PLoS One 12:e0181357. doi: 10.1371/journal.pone.0181357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rhee SY, Sankaran K, Varghese V, Winters MA, Hurt CB, Eron JJ, Parkin N, Holmes SP, Holodniy M, Shafer RW. 2016. HIV-1 protease, reverse transcriptase, and integrase variation. J Virol 90:6058–6070. doi: 10.1128/JVI.00495-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Woods CK, Brumme CJ, Liu TF, Chui CK, Chu AL, Wynhoven B, Hall TA, Trevino C, Shafer RW, Harrigan PR. 2012. Automating HIV drug resistance genotyping with RECall, a freely accessible sequence analysis tool. J Clin Microbiol 50:1936–1942. doi: 10.1128/JCM.06689-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Shafer RW, Hertogs K, Zolopa AR, Warford A, Bloor S, Betts BJ, Merigan TC, Harrigan R, Larder BA. 2001. High degree of interlaboratory reproducibility of human immunodeficiency virus type 1 protease and reverse transcriptase sequencing of plasma samples from heavily treated patients. J Clin Microbiol 39:1522–1529. doi: 10.1128/JCM.39.4.1522-1529.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Palmer S, Kearney M, Maldarelli F, Halvas EK, Bixby CJ, Bazmi H, Rock D, Falloon J, Davey RT Jr, Dewar RL, Metcalf JA, Hammer S, Mellors JW, Coffin JM. 2005. Multiple, linked human immunodeficiency virus type 1 drug resistance mutations in treatment-experienced patients are missed by standard genotype analysis. J Clin Microbiol 43:406–413. doi: 10.1128/JCM.43.1.406-413.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Schuurman R, Brambilla D, de Groot T, Huang D, Land S, Bremer J, Benders I, Boucher CA, ENVA Working Group. 2002. Underestimation of HIV type 1 drug resistance mutations: results from the ENVA-2 genotyping proficiency program. AIDS Res Hum Retroviruses 18:243–248. doi: 10.1089/088922202753472801. [DOI] [PubMed] [Google Scholar]
- 11.Church JD, Jones D, Flys T, Hoover D, Marlowe N, Chen S, Shi C, Eshleman JR, Guay LA, Jackson JB, Kumwenda N, Taha TE, Eshleman SH. 2006. Sensitivity of the ViroSeq HIV-1 genotyping system for detection of the K103N resistance mutation in HIV-1 subtypes A, C, and D. J Mol Diagn 8:430–432. doi: 10.2353/jmoldx.2006.050148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mohamed S, Penaranda G, Gonzalez D, Camus C, Khiri H, Boulme R, Sayada C, Philibert P, Olive D, Halfon P. 2014. Comparison of ultra-deep versus Sanger sequencing detection of minority mutations on the HIV-1 drug resistance interpretations after virological failure. AIDS 28:1315–1324. doi: 10.1097/QAD.0000000000000267. [DOI] [PubMed] [Google Scholar]
- 13.Pou C, Noguera-Julian M, Perez-Alvarez S, Garcia F, Delgado R, Dalmau D, Alvarez-Tejado M, Gonzalez D, Sayada C, Chueca N, Pulido F, Ibanez L, Rodriguez C, Casadella M, Santos JR, Ruiz L, Clotet B, Paredes R. 2014. Improved prediction of salvage antiretroviral therapy outcomes using ultrasensitive HIV-1 drug resistance testing. Clin Infect Dis 59:578–588. doi: 10.1093/cid/ciu287. [DOI] [PubMed] [Google Scholar]
- 14.Simen BB, Simons JF, Hullsiek KH, Novak RM, MacArthur RD, Baxter JD, Huang C, Lubeski C, Turenchalk GS, Braverman MS, Desany B, Rothberg JM, Egholm M, Kozal MJ, Terry Beirn Community Programs for Clinical Research on AIDS. 2009. Low-abundance drug-resistant viral variants in chronically HIV-infected, antiretroviral treatment-naive patients significantly impact treatment outcomes. J Infect Dis 199:693–701. doi: 10.1086/596736. [DOI] [PubMed] [Google Scholar]
- 15.Cozzi-Lepri A, Noguera-Julian M, Di Giallonardo F, Schuurman R, Daumer M, Aitken S, Ceccherini-Silberstein F, D'Arminio Monforte A, Geretti AM, Booth CL, Kaiser R, Michalik C, Jansen K, Masquelier B, Bellecave P, Kouyos RD, Castro E, Furrer H, Schultze A, Gunthard HF, Brun-Vezinet F, Paredes R, Metzner KJ, CHAIN Minority HIV-1 Variants Working Group. 2015. Low-frequency drug-resistant HIV-1 and risk of virological failure to first-line NNRTI-based ART: a multicohort European case-control study using centralized ultrasensitive 454 pyrosequencing. J Antimicrob Chemother 70:930–940. doi: 10.1093/jac/dku426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Le T, Chiarella J, Simen BB, Hanczaruk B, Egholm M, Landry ML, Dieckhaus K, Rosen MI, Kozal MJ. 2009. Low-abundance HIV drug-resistant viral variants in treatment-experienced persons correlate with historical antiretroviral use. PLoS One 4:e6079. doi: 10.1371/journal.pone.0006079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Varghese V, Shahriar R, Rhee SY, Liu T, Simen BB, Egholm M, Hanczaruk B, Blake LA, Gharizadeh B, Babrzadeh F, Bachmann MH, Fessel WJ, Shafer RW. 2009. Minority variants associated with transmitted and acquired HIV-1 nonnucleoside reverse transcriptase inhibitor resistance: implications for the use of second-generation nonnucleoside reverse transcriptase inhibitors. J Acquir Immune Defic Syndr 52:309–315. doi: 10.1097/QAI.0b013e3181bca669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Huber M, Metzner KJ, Geissberger FD, Shah C, Leemann C, Klimkait T, Boni J, Trkola A, Zagordi O. 2017. MinVar: a rapid and versatile tool for HIV-1 drug resistance genotyping by deep sequencing. J Virol Methods 240:7–13. doi: 10.1016/j.jviromet.2016.11.008. [DOI] [PubMed] [Google Scholar]
- 19.Trabaud MA, Icard V, Ramiere C, Tardy JC, Scholtes C, Andre P. 2017. Comparison of HIV-1 drug-resistance genotyping by ultra-deep sequencing and Sanger sequencing using clinical samples. J Med Virol 89:1912–1919. doi: 10.1002/jmv.24872. [DOI] [PubMed] [Google Scholar]
- 20.Learn GH Jr, Korber BT, Foley B, Hahn BH, Wolinsky SM, Mullins JI. 1996. Maintaining the integrity of human immunodeficiency virus sequence databases. J Virol 70:5720–5730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jabara CB, Jones CD, Roach J, Anderson JA, Swanstrom R. 2011. Accurate sampling and deep sequencing of the HIV-1 protease gene using a Primer ID. Proc Natl Acad Sci U S A 108:20166–20171. doi: 10.1073/pnas.1110064108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Varghese V, Wang E, Babrzadeh F, Bachmann MH, Shahriar R, Liu T, Mappala SJ, Gharizadeh B, Fessel WJ, Katzenstein D, Kassaye S, Shafer RW. 2010. Nucleic acid template and the risk of a PCR-induced HIV-1 drug resistance mutation. PLoS One 5:e10992. doi: 10.1371/journal.pone.0010992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Welkers MR, Jonges M, Jeeninga RE, Koopmans MP, de Jong MD. 2014. Improved detection of artifactual viral minority variants in high-throughput sequencing data. Front Microbiol 5:804. doi: 10.3389/fmicb.2014.00804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gianella S, Delport W, Pacold ME, Young JA, Choi JY, Little SJ, Richman DD, Kosakovsky Pond SL, Smith DM. 2011. Detection of minority resistance during early HIV-1 infection: natural variation and spurious detection rather than transmission and evolution of multiple viral variants. J Virol 85:8359–8367. doi: 10.1128/JVI.02582-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Keys JR, Zhou S, Anderson JA, Eron JJ Jr, Rackoff LA, Jabara C, Swanstrom R. 2015. Primer ID informs next-generation sequencing platforms and reveals preexisting drug resistance mutations in the HIV-1 reverse transcriptase coding domain. AIDS Res Hum Retroviruses 31:658–668. doi: 10.1089/aid.2014.0031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhou S, Jones C, Mieczkowski P, Swanstrom R. 2015. Primer ID validates template sampling depth and greatly reduces the error rate of next-generation sequencing of HIV-1 genomic RNA populations. J Virol 89:8540–8555. doi: 10.1128/JVI.00522-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Clutter DS, Zhou S, Varghese V, Rhee SY, Pinsky BA, Fessel WJ, Klein DB, Spielvogel E, Holmes SP, Hurley LB, Silverberg MJ, Swanstrom R, Shafer RW. 2017. Prevalence of drug-resistant minority variants in untreated HIV-1-infected individuals with and those without transmitted drug resistance detected by Sanger sequencing. J Infect Dis 216:387–391. doi: 10.1093/infdis/jix338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dauwe K, Staelens D, Vancoillie L, Mortier V, Verhofstede C. 2016. Deep sequencing of HIV-1 RNA and DNA in newly diagnosed patients with baseline drug resistance showed no indications for hidden resistance and is biased by strong interference of hypermutation. J Clin Microbiol 54:1605–1615. doi: 10.1128/JCM.00030-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Noguera-Julian M, Cozzi-Lepri A, Di Giallonardo F, Schuurman R, Daumer M, Aitken S, Ceccherini-Silberstein F, D'Arminio Monforte A, Geretti AM, Booth CL, Kaiser R, Michalik C, Jansen K, Masquelier B, Bellecave P, Kouyos RD, Castro E, Furrer H, Schultze A, Gunthard HF, Brun-Vezinet F, Metzner KJ, Paredes R, CHAIN Minority HIV-1 Variants Working Group. 2016. Contribution of APOBEC3G/F activity to the development of low-abundance drug-resistant human immunodeficiency virus type 1 variants. Clin Microbiol Infect 22:191–200. doi: 10.1016/j.cmi.2015.10.004. [DOI] [PubMed] [Google Scholar]
- 30.Delviks-Frankenberry KA, Nikolaitchik OA, Burdick RC, Gorelick RJ, Keele BF, Hu WS, Pathak VK. 2016. Minimal contribution of APOBEC3-induced G-to-A hypermutation to HIV-1 recombination and genetic variation. PLoS Pathog 12:e1005646. doi: 10.1371/journal.ppat.1005646. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Babrzadeh F, Varghese V, Pacold M, Liu TF, Nyren P, Schiffer C, Fessel WJ, Shafer RW. 2013. Collinearity of protease mutations in HIV-1 samples with high-level protease inhibitor class resistance. J Antimicrob Chemother 68:414–418. doi: 10.1093/jac/dks409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Castro H, Pillay D, Cane P, Asboe D, Cambiano V, Phillips A, Dunn DT, UK Collaborative Group on HIV Drug Resistance. 2013. Persistence of HIV-1 transmitted drug resistance mutations. J Infect Dis 208:1459–1463. doi: 10.1093/infdis/jit345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Jain V, Sucupira MC, Bacchetti P, Hartogensis W, Diaz RS, Kallas EG, Janini LM, Liegler T, Pilcher CD, Grant RM, Cortes R, Deeks SG, Hecht FM. 2011. Differential persistence of transmitted HIV-1 drug resistance mutation classes. J Infect Dis 203:1174–1181. doi: 10.1093/infdis/jiq167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Pingen M, Wensing AM, Fransen K, De Bel A, de Jong D, Hoepelman AI, Magiorkinis E, Paraskevis D, Lunar MM, Poljak M, Nijhuis M, Boucher CA, SPREAD Programme. 2014. Persistence of frequently transmitted drug-resistant HIV-1 variants can be explained by high viral replication capacity. Retrovirology 11:105. doi: 10.1186/s12977-014-0105-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Li JZ. 2014. HIV-1 drug-resistant minority variants: sweating the small stuff. J Infect Dis 209:639–641. doi: 10.1093/infdis/jit656. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Kyeyune F, Gibson RM, Nankya I, Venner C, Metha S, Akao J, Ndashimye E, Kityo CM, Salata RA, Mugyenyi P, Arts EJ, Quinones-Mateu ME. 2016. Low-frequency drug resistance in HIV-infected Ugandans on antiretroviral treatment is associated with regimen failure. Antimicrob Agents Chemother 60:3380–3397. doi: 10.1128/AAC.00038-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Charpentier C, Lee GQ, Rodriguez C, Visseaux B, Storto A, Fagard C, Molina JM, Katlama C, Yazdanpanah Y, Harrigan PR, Descamps D. 2015. Highly frequent HIV-1 minority resistant variants at baseline of the ANRS 139 TRIO trial had a limited impact on virological response. J Antimicrob Chemother 70:2090–2096. doi: 10.1093/jac/dkv048s. [DOI] [PubMed] [Google Scholar]
- 38.Westby M, Lewis M, Whitcomb J, Youle M, Pozniak AL, James IT, Jenkins TM, Perros M, van der Ryst E. 2006. Emergence of CXCR4-using human immunodeficiency virus type 1 (HIV-1) variants in a minority of HIV-1-infected patients following treatment with the CCR5 antagonist maraviroc is from a pretreatment CXCR4-using virus reservoir. J Virol 80:4909–4920. doi: 10.1128/JVI.80.10.4909-4920.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Li JZ, Paredes R, Ribaudo HJ, Svarovskaia ES, Metzner KJ, Kozal MJ, Hullsiek KH, Balduin M, Jakobsen MR, Geretti AM, Thiebaut R, Ostergaard L, Masquelier B, Johnson JA, Miller MD, Kuritzkes DR. 2011. Low-frequency HIV-1 drug resistance mutations and risk of NNRTI-based antiretroviral treatment failure: a systematic review and pooled analysis. JAMA 305:1327–1335. doi: 10.1001/jama.2011.375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Boltz VF, Bao Y, Lockman S, Halvas EK, Kearney MF, McIntyre JA, Schooley RT, Hughes MD, Coffin JM, Mellors JW, OCTANE/A5208 Team. 2014. Low-frequency nevirapine (NVP)-resistant HIV-1 variants are not associated with failure of antiretroviral therapy in women without prior exposure to single-dose NVP. J Infect Dis 209:703–710. doi: 10.1093/infdis/jit635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Avila-Rios S, Garcia-Morales C, Matias-Florentino M, Romero-Mora KA, Tapia-Trejo D, Quiroz-Morales VS, Reyes-Gopar H, Ji H, Sandstrom P, Casillas-Rodriguez J, Sierra-Madero J, Leon-Juarez EA, Valenzuela-Lara M, Magis-Rodriguez C, Uribe-Zuniga P, Reyes-Teran G, HIVDR MexNet Group. 2016. Pretreatment HIV-drug resistance in Mexico and its impact on the effectiveness of first-line antiretroviral therapy: a nationally representative 2015 WHO survey. Lancet HIV 3:e579–e591. doi: 10.1016/S2352-3018(16)30119-9. [DOI] [PubMed] [Google Scholar]
- 42.Perrier M, Visseaux B, Landman R, Joly V, Todesco E, Yazdanpanah Y, Calvez V, Marcelin AG, Descamps D, Charpentier C. 2017. No impact of HIV-1 protease minority resistant variants on the virological response to a first-line PI-based regimen containing darunavir or atazanavir. J Antimicrob Chemother 73:173–176. doi: 10.1093/jac/dkx366. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.