Auckland in lockdown, and the rest of the country at level 2 after four cases of community transmission are detected.
Scientists have turned the complex tale of Auckland's August Covid-19 cluster into a fascinating interactive, threading together its cases with genomic data. Science reporter Jamie Morton talked to Dr James Hadfield, a Wanaka-based phylogeneticist with Seattle's Bedford Lab, about what we've learned - and the three likely explanations for the mystery outbreak.
Genomic sequencing helps scientists and public health officials identify lineages of each positive case and connect them to other local cases, along with what's circulating in other countries. How did the genomic profile of the Auckland August cluster differ to that of New Zealand's previous clusters?
Revisiting our first series of outbreaks, between February and June, we know there were more than 250 separate introductions of Sars-CoV-2 into Aotearoa New Zealand, which represented "nearly all of the genomic diversity present in the global viral population".
Scientists have created a fascinating interactive tale of how Covid-19 returned to New Zealand in August. Image / Nextstrain
The Auckland August cluster belongs to a lineage that is in multiple continents around around the world. We have seen this lineage only once before in New Zealand, in a pair of cases in mid-April who were in managed isolation in Auckland.
Given the timing, the epidemiological data, and the worldwide prevalence of the lineage, we think it highly probable that the Auckland August cluster is a new import from overseas.
How can we tell that Auckland's recent outbreak came from a single viral source?
Sequencing the genomes of many different cases from the outbreak shows they all grouped together, with no worldwide isolates involved.
This is what we'd expect for a cluster that has spread within New Zealand and originated from a single viral source. The exception to this was the Rydges hotel case.
Can you tell me about Nextstrain - the global platform you're contributing to - and the interactive narrative you've created here? What does it show us?
Nextstrain has been around for around four years and is an academic project that uses genome sequencing of pathogens such as Covid-19 to produce rapid analyses.
The aim is that these results are "actionable" — or that they're useful for informing public health responses.
We disseminate results freely through nextstrain.org for everyone to interact with.
Understanding these analyses can be tricky without a background in phylogenetics, and so I've been exploring ways we can better communicate the insights genomics give us.
Narratives allow scientists to write reports where each section of text is shown next to a specific visualisation of the data.
In other words, the scientist can guide the viewer through the data and convey the insights they glean from it.
People queue outside Countdown Hobsonville on August 12, soon after the outbreak was announced. Photo / Dean Purcell
We wrote this particular narrative for Kiwis to read, with the aim of better communicating the data and what it does - and doesn't - tell us about the current situation.
I hope people find this medium accessible.
The genomic analysis suggests three possible sources for the outbreak: a border incursion from overseas, namely from the UK or Ecuador; undetected community transmission; and the virus arriving via foreign goods. Can you explain each of these - and which might be the most likely?
The first, and most likely, hypothesis is a border incursion from overseas - that an international arrival brought back the virus and subsequently transmitted it to the community.
Finding a recent case in managed isolation and quarantine (MIQ) or elsewhere on the border that matched this cluster lineage would be strong evidence for the border incursion scenario.
The August outbreak sent a steady stream of cars to the Covid-19 testing site in Newbury St, Otara. Photo / Alex Burton
We haven't got any such direct evidence - but this doesn't mean it didn't happen. Although we have attempted to sequence all the positive cases identified at the border, around 40 per cent contain too little or too degraded genetic material (RNA) to obtain a full genome sequence.
Additionally, the swab test for Covid-19 is known to miss a proportion of cases, so even the MIQ testing regime, with tests on day three and 12, would be expected to miss about 4 per cent of true-positive cases.
Due to how genetically similar the Auckland Cluster genome is to others around the world, it's hard to use those international samples to pinpoint specific geographical sources with any certainty.
For instance, the closest samples we have are from Ecuador, but there were no returnees from Ecuador into New Zealand from the start of June until the end of July - when the first Auckland cluster case developed symptoms - so that's unlikely to be the route.
The second hypothesis is that the source was a case of undetected community transmission from our previous outbreaks.
One New Zealand isolate is genetically similar to the current Auckland cluster, which came from a case who was known to be infected before arrival in New Zealand and was immediately quarantined in Auckland in April.
Given the mutations between the April case and the August cluster, and the timeframe involved, our analysis estimates that there is less than 1 per cent chance this case was the source of the outbreak.
Extensive testing and contact tracing determined the earliest case found to date was an Americold coolstore worker who first showed symptoms on July 31, and the initial spread of the cluster centred around this coolstore, which imports frozen goods.
Thus, the third hypothesis is that the virus may have been imported on packaging material, where it could have survived in low temperature conditions, and then gone on to infect a worker at the coolstore.
This hypothesis is given further credence by the possible genomic link to Ecuador, since viral particles have been found in China on frozen shrimp packaging from Ecuador.
However, no shipments from Ecuador were received by the coolstore in question.
Our analysis shows that the estimated mutation rate on the branch leading to the cluster is not a lot smaller than elsewhere in the tree, lending little weight to the possibility that the virus lay dormant on packing material for a long period of time.
Environmental testing of the coolstore facility on August 15 found one weak positive sample on packaging, which was considered consistent with contamination.
Confirming or excluding fomite transmission is very hard to prove in non-laboratory or un-controlled settings, but thus far it seems that fomite transmission is not very common.
The narrative presents all of these in more detail alongside interactive visualisations of the data.
Is it fair to assume we'll never learn exactly what that source was?
It's quite possible we'll never know — each situation is different.
We've seen examples since then - the Rydges case and the recent "Christchurch returnee cases" - where the data was more conclusive in identifying the source.
That Rydges Hotel case really proved an interesting outlier. What did this tell us about the value of rapid genome sequencing - and also that random pop-ups like this can happen?
Within two days of the positive test result, the sequencing data showed that the case was not part of the wider Auckland cluster, but rather that it was very similar to a returnee who had stayed at the Rydges before testing positive.
A sole case at Auckland's Rydges Hotel, not linked to the rest of the August outbreak, proved a mystery within a wider mystery. Photo / Dean Purcell.
This stopped contact tracers spending valuable time searching for a - non-existent - link between this case and the then-ongoing Auckland cluster.
The analysis also makes the point that our understanding of the genomic picture might change. Why is this?
Tracking outbreaks such as these relies on comparing genomes from the Auckland cluster with everything else that's available.
If we are able to get more genomes - for instance, from July MIQ cases, which we have been unable to sequence thus far - then that data may give us more confidence in one of the hypothesis we present here.
A central idea behind nextstrain is that as we get more data we can continue to update the analysis and our findings.
