SARS-CoV-2 outbreak investigation: from sample to response in a day
Áine O'Toole (Institute of Evolutionary Biology, University of Edinburgh, United Kingdom) started her talk by going back to the very beginning of the SARS-CoV-2 outbreak: when 27 cases of viral pneumonia were reported in December 2019. By the 1st January 2020, the virus had been identified as a novel coronavirus, and by the 10th, the first whole viral genome sequence had been shared. By November 2020, the number of COVID-19 cases had reached 48 million across the globe. However, Áine highlighted, ‘we’ve also seen an unprecedented global genomics response’, with 187,000 submissions now uploaded to GISAID. With the time and effort invested in SARS-CoV-2 genomics, Áine stressed how crucial it is for useful, actionable information to be fed back as soon as possible.
To this end, Áine and her colleagues have used nanopore sequencing to develop an end-to-end system that can go from sample to useable, interpretable information within a single day. At the Royal Infirmary, Edinburgh, they are routinely using genomics to investigate SARS-CoV-2 outbreaks in real time. First, RNA is extracted, then RT-PCR & library prep methods including the ARTIC protocol are used to prepare samples for sequencing. From raw sample to sequencing, preparation takes 9 hours; libraries are then sequenced on MinION, or GridION, providing higher capacity to sequence many samples. Basecalling and demultiplexing is performed in real-time, and the coverage for each sample is assessed as the run progresses via Rampart, so that they ‘only need to sequence for as long as it takes to generate enough read coverage’, which can require very little time. The flow cells are also nuclease washed and re-used. Áine noted that the ARTIC protocol and Rampart have been used across the world during the COVID-19 pandemic. The ARTIC MinION pipeline is employed for data analysis; Áine described how, with the recent addition of Medaka, ‘consensus genomes can be generated really very rapidly’.
Having described how she and her team rapidly generate SARS-CoV-2 consensus sequences, Áine went on to discuss her main focus for this year: how to use this data to generate actionable information, for same-day outbreak investigation. With the goal of making this process as straightforward as possible, she and her colleague, Verity Hill, developed civet: ‘cluster investigation & virus epidemiology tool’. This runs automated phylogenetic analysis, then generates a summary report featuring a range of figures to help inform an outbreak. Civet is being used routinely by researchers across the UK to investigate local COVID-19 outbreaks, rule out cases of transmission, and supplement contact tracing efforts. The tool works by taking SARS-CoV-2 sequences – ‘hot off the GridION or MinION’ - and finding their closest sequence in the global phylogeny and local catchment area. It then collapses this section of the tree to display only relevant information; these collapsed trees are very quick to re-run with newly added sequences included. The summary report produced features available metadata for sequences of interest, spatial context highlighting the diversity of strains circulating in the surrounding area, and timelines for each case to aid investigations. It also runs snipit: a tool which calls variants in each sample and helps highlight identical sequences. Áine noted that, whilst it is tempting to use identical sequences as evidence of transmission, they have been very careful to provide guidance on interpretation: whilst they can’t answer ‘what is a transmission?’, they can ask ‘what isn’t?’.
Áine highlighted that this method of investigation is dependent on the availability of many up-to-date SARS-CoV-2 genome sequences, which relies on the hard work of the COVID-19 Genomics UK Consortium (COG-UK), a highly collaborative effort that has sequenced over 94,000 viruses. COG has centres across the UK, enabling decentralised sequencing and bioinformatics. This local data is uploaded to a central server, QC is performed, and then phylogenetic analysis can take place. With this combination of decentralised sequencing and rapid data sharing, civet can be run once every 24 hours and the information disseminated back out, providing an up-to-date, global phylogenetic tree for SARS-CoV-2.