P09. Multiplexed next-generation sequencing and de novo assembly to obtain near full-length HIV-1 genome from plasma virus

Shambhu G. Aralaguppe1, Abu Bakar Siddik1, Ashokkumar Manickam2, Wondwossen Amogne3, Luke Elizabeth Hanna2, Anders Sonnerborg1, Ujjwal Neogi1
Affiliates: 1Karolinska Institute, Stockholm, Sweden 2National Institute for Research in Tuberculosis, Indian Council of Medical Research, Chennai, India 3Addis Ababa University, Addis Ababa, Ethiopia

View posterpresentation

Background
One of the most prominent characteristics of HIV is the incredible genetic diversity that gave rise to several subtypes and recombinant forms. Near full-length HIV-1 genome sequence (HIV-NFLG) obtained directly from plasma viruses are scarce, but contributes to high-resolution molecular epidemiological information. Analyzing the HIV-NFLG facilitates new understanding into the diversity of virus population dynamics at individual or population level. High genetic diversity not only presents major challenges for vaccine development but also foremost risk for effective diagnostics and monitoring assays. In this study we developed a simple but high-throughput next generation sequencing (NGS) protocol for HIV-NFLG using clinical specimens and validated the method against an external quality control (EQC) panel.

Methods
Clinical specimens (n = 105) were obtained from three cohorts from two highly conserved HIV-1C epidemics (India and Ethiopia) and one diverse epidemic (Sweden). HIV-NFLG was performed amplifying the HIV-genome in two fragments (F1-Gag-to-vpu and F2-Tat-to-3LTR). NGS was performed in a subset of samples (n=45) in the Illumina HiSeq2500 after multiplexing 24 samples, followed by de novo assembly in Iterative Virus Assembler (IVA) or VICUNA. Additionally an EQC panel (n = 10) was used to validate the protocol. Subtyping was carried out using Rega v3 and COMET HIV-1 followed by HIV-1 jpHMM.

Results
Amongst 105 plasma samples used, F1 and F2 fragments were amplified from 95 samples giving the success rate of 90% amplification. All the 45 patient samples and 10 EQC panel attempted for NGS were successful. De novo assembly was successful for 49 samples using IVA. For the other 6 samples where IVA failed to generate the NFLG contig, VICUNA was used. The mean coverage per position was >10,000x. The mean error for mutations for the EQC panel viruses and the MJ4 plasmid were <1%. This indicates very low level of PCR induced errors and errors due to NGS. HIV-1 subtyping by automated tools and maximum likelihood phylogenetic analysis identified 87% (39/45) as HIV-1C, 8% (4/45) as HIV-1B and the remaining (2/45) samples were identified as A1C recombinants.

Conclusions
Here we demonstrate the practicality of a high-throughput NGS protocol to assemble near full-length HIV-1 genomes from plasma virus that is frequently used as the source for most routine HIV-1 genotypic tests. The major advantage of the HIV-NFL-multiplexed NGS is the broad applicability. Here we applied the method in three distinct cohorts and obtained high success in all three. The method is also less labor-intensive and simple in both wet (PCR/NGS) and dry (bioinformatics) laboratory setups. The input plasma sample volume is only 140 μl, thus it can be combined with the existing routine genotypic resistance testing without any additional blood specimen. Finally, the method is validated against the EQC panel, thus quality assured for large-scale molecular epidemiological surveillance studies, even in LMICs.