Investigating an agnostic metagenomic virus discovery tool and evaluating the performance of Ion Torrent versus Oxford Nanopore Technology sequencing

Abstract

Introduction and motivation

Emerging and re-emerging viruses are a global health concern. Historically, such viruses have been identified through targeted approaches, which has been the gold standard in diagnosis. Unfortunately, these approaches are limited to known pathogens and thus inadequate for the monitoring of “unknown” viruses. Therefore, there is a need for a technology that can identify known, novel and unexpected viruses.

Virus discovery using copy deoxyribonucleic acid (cDNA) and amplified fragment length polymorphism (AFLP) (also called VIDISCA) is a metagenomics approach that non-specifically enriches for viral sequences and then uses next generation sequencing (NGS) for virus discovery. VIDISCA has been used for the discovery of a several previously unknown viruses, including human coronavirus (HCoV) NL63 and Ntwetwe virus.

VIDISCA has previously relied on Ion Torrent NGS, and we hypothesised that Oxford Nanopore Technology (ONT), which is easier to use, could be as effective as Ion Torrent sequencing. Therefore, we developed and validated VIDISCA with ONT and this compared to Ion Torrent NGS.

Methodology

A library made-up of different sample types was split into two identical libraries to sequence on the ONT and Ion Torrent platforms. The samples contained known viral pathogens which were blinded and treated as “unknown” samples for viral discovery. cDNA was synthesized with random primers from nucleic acid extracts that were depleted of rRNA complementary sequences. After second strand synthesis, DNA was digested with a restriction enzyme coupled with adaptor ligation, which served as priming sites for PCR enrichment. Templates are then sequenced, and output files were mapped against protein databases including human, mammal, insect, tick and avian virus sequences.

Findings

The two NGS platforms performed similarly in identifying the pathogens in the prepared library that contained Human Mastadenovirus (HAdV), Human Immunodeficiency virus (HIV), Hepatitis E virus (HEV), HCoV-NL63 and a virus free sample. Neither the ONT nor Ion Torrent platforms sequenced the HAdV. Both platforms also missed HEV in one sample. All other viruses were successfully identified by both platforms. Overall, ONT performed similarly to Ion Torrent.

Conclusion and relevance

VIDISCA performed on Ion Torrent and ONT sequencing platforms has displayed similar Results. ONT sequencing has the advantage of a more rapid turnaround and low sequencing footprint with a similar cost estimation.

Future plans include the testing of residual patient samples where some viruses are often undiagnosed (such as cerebrospinal fluid) as well as sentinel species (such as rats and shrews) as a part of pandemic preparedness and novel virus discovery. If necessary, additional work, such as genome primer walking, will be conducted for full-genome characterization of newly discovered viruses identified with VIDISCA-NGS.