Interview: The frontlines of pathogen genomic surveillance with rapid nanopore sequencing
During a recent webinar, we heard from Joana Isidro, from National Institute of Health Doutor Ricardo Jorge, Alex Shaw, from Imperial College London, and Jonathan Edgeworth, from Guy’s & St. Thomas’ NHS Foundation Trust & Kings College London. They discussed their work on genomic surveillance, with an emphasis on rapid metagenomics methods using nanopore sequencing. During the webinar, we weren't able to answer all of the questions during the live Q&A, so we caught up with the speakers after the event.
You can also watch the full broadcast of the webinar ‘The frontlines of pathogen genomic surveillance with rapid nanopore sequencing webinar’ on demand here.
What do you do with the rest of the sequences that are not monkeypox? 0.5% is a very low amount, isn't this a waste of resources? Is it there a way to enrich for viral genomes before library prep?
Joana: Our approach for this first sequencing run was shotgun metagenomics and therefore the sequencing of host DNA is always expected. In fact, no other targeted approaches, such as the whole-genome amplification performed for SARS-CoV-2 with interspersed amplicons, are available yet for this virus — especially due to the large MPXV genome size. However, we are optimising the DNA extraction protocols with a pre-treatment of sonication and a cocktail of DNAses/RNAses with the aim of depleting host DNA in the sample, and so increasing the proportion of viral DNA prior to the DNA extraction protocol. This is expected to lead to a higher percentage of viral reads per sample when using a metagenomics approach.
Oxford Nanopore: In addition to metagenomics protocols, groups have recently released PCR-based amplicon tiling approaches for Monkeypox virus WGS such as https://www.protocols.io/view/monkeypox-virus-whole-genome-sequencing-using-comb-ccc7sszn.html
Given that you didn’t utilise human depletion or enrichment with PCR, how long did you need to sequence to get enough coverage for your first draft genome?
Joana: The sequencing run length was about 18 hours and we monitored the reads mapping against a MPXV reference genome during the run to infer the approximate coverage we had at each point.
Great presentation, and good to see that reference assembly of reads helped provide the needed information and a West Africa linkage. Can you comment on the limitations of reference assembly (e.g., one versus many reference genomes) and advantages of an additional de novo assembly?
Joana: Our approach started with a de novo assembly that was used to identify the closest reference genome sequence available at the time in public repositories. We then mapped the reads against this reference genome. But in parallel, we also inspected the contigs from the de novo assembly to verify any structural differences or large indels that could be present in our sample. Both approaches, together with visual inspection of mutations in mapped reads, were applied to curate the final genome sequence.
Is INSaFLU capable of analysing raw bacterial data and genome assembly?
Joana: INSaFLU is tailored to deal with amplicon-derived sequencing data and small genomes, having some modules tailored to viruses, namely influenza, SARS-CoV-2, and now also Monkeypox.
Can we use the same protocol for shotgun metagenomics of influenza virus?
Joana: This rapid barcoding protocol can be used for genome sequencing of influenza virus from clinical research samples, provided that viral RNA is reverse-transcribed to obtain double-stranded cDNA. Nonetheless, it is worth noting that, contrarily to MPXV, there are amplicon-based protocols that allow the simultaneous amplification of the eight viral segments of influenza prior to sequencing. With the latter approach, one can expect a high success rate with high depth of coverage, given that host contamination is minimised.
How long did you sequence for to get 7-fold depth of coverage of the MPX genome? Is there a minimum depth recommended for consensus sequence generation?
Joana: The sequencing run length was about 18 hours, but the success of the sequencing run, the number of reads generated, and the successful sequencing of the whole virus depends on multiple factors; for example, the total DNA concentration of your sample and the proportion of MPXV in the sample.
To identify the 50 SNPs in your assembly did you use any additional targeted sequencing runs for error correction?
Joana: No, we visually inspected the mapping of reads throughout the whole 200 kb genome to confirm all the variant positions.
Can we use the same protocol for African swine fever virus genome sequencing?
Joana: We do not have experience with African swine fever virus genome sequencing but as it also contains a large dsDNA genome, this protocol might be used to attempt whole virus genome sequencing. However, it is important to note that the success of shotgun metagenomic approaches depends on multiple factors, including the proportion of the virus in the sample, the total concentration of the extracted nucleic acids, nucleic acid quality and integrity, sequencing depth, etc.
Oxford Nanopore: for more information on this topic, you can watch ‘The other deadly virus: tiled amplicon sequencing and assembly of African swine fever virus’ presented by Amanda Warr and Christine Tait-Burkard and find a protocol for PCR tiling of the African swine fever virus here.
What is the maximum Ct detection limit with this protocol?
Joana: A lower Ct value maximizes the likelihood of success but does not guarantee it, since the success will depend on multiple factors. The first genome was obtained from a sample with Ct=17 and we have since attempted sequencing mostly in samples with Ct values below 25. We have also started applying a pre-treatment (sonication and a DNAse/RNAse cocktail) to the samples prior to sequencing to deplete host DNA and maximise the proportion of MPXV in the sample.
Was there any special consideration for sample collection, preservation, or extraction to facilitate full genome poliovirus amplification?
Alex: The faecal samples are stored at fridge temperature and ideally taken to the lab for storage at -20°C within three days. They are not stored with any additional preserving agents. We find that for RNA extractions, bead-based kits are preferable, and have had very poor results using total nucleic acid extracted following bead-beating homogenisation.
Thank you for the great speech. What is the sensitivity of Direct Detection by Nanopore Sequencing (DDNS)? What is the lowest working vrial load in this assay? How long is the amplicon in DDNS assay?
Alex: We have successfully amplified as few as 2 RNA copies per μL, with the corresponding viral load depending on the efficiency of the RNA extraction. The DDNS amplicon is approximately 1200 bp, fully covering the VP1 region of the capsid.
How is it that the cell line is only more susceptible to Polio and not other pathogens? Does this create bias within that assay? Does your new workflow then eliminate that bias?
Alex: The poliovirus specific cell line is L20B. The L20B line are mouse L cells transfected with the human poliovirus receptor (Mendelsohn et al., 1989). This does allow selection for poliovirus, which is the intention of the cell-culture assay, and we achieve similar selection by using the poliovirus specific VP1 primers Y7 and Q8.
For Dr. Alex Shaw, what type of concentration method is recommended for environmental samples (surface waters) for a successful sequencing?
Alex: This can vary greatly depending on your target organism. For poliovirus, samples tend to be processed by ultracentrifugation or two-phase separation.
Thank you for giving such as great talk. I am wondering if there multiple viral quasispecies in your experiments? If yes, could viral population structure change in DDNS?
Alex: We do find mixture of viruses in our samples, particularly in the environmental (sewage) samples. We have managed to separate out these closely related strains and are updating the software used as standard in our analysis pipeline to perform this.