Towards Next-Generation Sequencing for HIV-1 Drug Resistance Testing in a Clinical Setting

Performance of the Applied Biosystems HIV-1 Genotyping Kit with Integrase

HIV genotyping is used to assess HIV susceptibility to antiretroviral drugs. The Applied Biosystems HIV-1 Genotyping Kit with Integrase (AB kit, Thermo Fisher Scientific) detects resistance-associated mutations (RAMs) in HIV protease (PR), reverse transcriptase (RT), and integrase (IN). We compared results from the AB kit with results obtained previously with the ViroSeq HIV-1 Genotyping System. DNA amplicons from the AB kit were also analyzed using next-generation sequencing (NGS). HIV RNA was extracted using the MagNA Pure 24 instrument (Roche Diagnostics; 96 plasma samples, HIV subtype B, viral load range: 530-737,741 copies/mL). FASTA files were generated from AB kit data using Exatype (Hyrax Biosciences). DNA amplicons from the AB kit were also analyzed by NGS using the Nextera XT kit (Illumina). Drug resistance was predicted using the Stanford HIV Drug Resistance Database. The mean genetic distance for sequences from ViroSeq and the AB kit was 0.02% for PR/RT and 0.04% for IN; 103 major RAMs were detected by both methods. Four additional major RAMs were detected by the AB kit only. These four major RAMs were also detected by NGS (detected in 18.1%-38.2% of NGS reads). NGS detected 27 major RAMs that were not detected with either of the Sanger sequencing-based kits. All major RAMs detected with ViroSeq were detected with the AB kit; additional RAMs were detected with the AB kit only. DNA amplicons from the AB kit can be used for NGS for more sensitive detection of RAMs.

The performance of homopolymer detection using dichromatic and tetrachromatic fluorogenic next-generation sequencing platforms

OBJECTIVES

Homopolymer (HP) sequencing is error-prone in next-generation sequencing (NGS) assays, and may induce false insertion/deletions and substitutions. This study aimed to evaluate the performance of dichromatic and tetrachromatic fluorogenic NGS platforms when sequencing homopolymeric regions.

RESULTS

A HP-containing plasmid was constructed and diluted to serial frequencies (3%, 10%, 30%, 60%) to determine the performance of an MGISEQ-2000, MGISEQ-200, and NextSeq 2000 in HP sequencing. An evident negative correlation was observed between the detected frequencies of four nucleotide HPs and the HP length. Significantly decreased rates (P < 0.01) were found in all 8-mer HPs in all three NGS systems at all four expected frequencies, except in the NextSeq 2000 at 3%. With the application of a unique molecular identifier (UMI) pipeline, there were no differences between the detected frequencies of any HPs and the expected frequencies, except for poly-G 8-mers using the MGI 200 platform. UMIs improved the performance of all three NGS platforms in HP sequencing.

CONCLUSIONS

We first constructed an HP-containing plasmid based on an EGFR gene backbone to evaluate the performance of NGS platforms when sequencing homopolymeric regions. A highly comparable performance was observed between the MGISEQ-2000 and NextSeq 2000, and introducing UMIs is a promising approach to improve the performance of NGS platforms in sequencing homopolymeric regions.

Evaluation of the ABL NGS assay for HIV-1 drug resistance testing

HIV evolution and variability around the world requires special monitoring of the viral strains in infected people. High-throughput HIV sequencing and drug resistance testing techniques have become routinely available over the last few years. We conducted a study to assess the new CE-marked ABL NGS HIV genotyping assay on an Illumina® platform, to compare the results (the detection of resistance associated mutations (RAMs) detected in the three main targets: reverse transcriptase, protease, and integrase) with those produced by three Sanger-based assays, and to compare the assays' respective costs. For the 10 samples and a 20 % sensitivity threshold for the NGS technology, the percent agreement between the four assays ranged from 99.5 % to 100 %. We detected 4 more and 10 more RAMs of interest when we lowered the NGS assay's threshold to 10 % and 3 %, respectively. At a threshold of 3 %, the antiretroviral sensitivity interpretation algorithm (for protease inhibitors) was modified for only two patients. The NGS assay's unit cost fell rapidly as the number of samples per run increased. Compared with Sanger sequencing, the ABL NGS HIV genotyping assay is just as robust and somewhat more expensive but opens up interesting multiplexing perspectives for virology laboratories.