SARS-CoV-2 genomics from research to public health
For their first recording with all three hosts together in person, Lee Katz, Andrew Page and Nabil-Fareed Alikhan welcome Torsten Seemann. They compare their pandemic work, from investigating transmission and reinfection to building sequencing services and communicating results to public health teams. The discussion shows why useful genomic surveillance needs reliable laboratory processes, epidemiological context and established relationships—not just sequence data.
Listen to Episode 95: Lost in Translation with Torsten Seemann on SoundCloud
In this episode
Flights, fieldwork and laboratory lessons
Torsten describes how Australia’s returning travellers provided introductions of SARS-CoV-2 despite closed borders. Travel histories and hotel-quarantine records gave investigators detailed epidemiological information. In some analyses, genomic evidence supported transmission during flights: passengers arriving from different countries had boarded the same aircraft at a hub such as Singapore. He thinks this led to some policy on masks, passenger separation and air filtration.
The panel wonders whether repeated rapid antigen tests or rapid PCR could track infection during a long flight. That leads to a discussion of portable laboratories, MinION fieldwork and improvised centrifuges. Torsten recalls a seven-day molecular biology course involving extraction, cloning, expression and capillary sequencing, plus several mishaps that earned him a destruction award. Lee’s experience learning PFGE at CDC similarly helped him understand laboratory colleagues’ work. Both accounts emphasise appreciating practical processes and where errors can enter.
When genomic results trigger restrictions
Moving from research into public health meant adjusting to stricter controls and procedures. Torsten explains that this initially felt frustrating, but the consequences justified the caution. His team’s genomic results helped uncover a major hotel-quarantine outbreak and contributed to a residential tower being locked down, affecting people confined to their building for weeks. Contamination or a sample swap could therefore have serious consequences.
Trust had been built before COVID through work on foodborne pathogens and consultations with epidemiologists and health officials. During the pandemic, bioinformaticians attended important meetings alongside genomic epidemiologists and medical microbiologists. Their role included explaining technical limitations: missing SARS-CoV-2 data, for example, could change a sample’s position in a phylogenetic tree. Torsten describes bioinformaticians becoming visible participants in decision-making rather than remaining behind the scenes.
Reinfection and the limits of consensus genomes
Lee recalls being asked early in the response whether people could be reinfected. He considered an all-versus-all genomic comparison of samples from successive infections and found an unnamed preprint involving Torsten that used the approach he had in mind. Torsten discusses comparisons of within-host and between-patient diversity, while acknowledging questionable data and limited time.
Consensus sequences alone could miss a mixture or a minor allele in an initial sample that later became dominant. Torsten says revisiting raw FASTQ data would have helped investigate that possibility. Publicly deposited reads offer opportunities for later studies, but only if metadata can connect samples from the same patient. The discussion also covers the SIREN project sampling healthcare workers and thresholds such as 90 days between positive tests, intended to help distinguish reinfection from a continuing infection.
Turning research laboratories into COG-UK production
Andrew describes existing bacterial-genomics networks being rapidly drawn into COG-UK, with Nick Loman and Sharon Peacock among those mobilising collaborators. Local hospital connections and a shared campus with a regional viral diagnostic unit helped his team obtain samples. Sequencing expanded into bioinformatics analysis, national surveillance and international studies, while a research laboratory adopted the regular output and quality controls expected in public health.
Josh Quick’s ARTIC protocol helped the team get started. Andrew recalls reaching production sequencing within about eight days of receiving the go-ahead, with ethics and contracts expedited. Sample collection began on 8 April 2020, and sequencing continued every week, including Christmas, until mid-April 2022. The panel also credits protocols.io as a useful resource for sharing protocols across countries.
Clinical identifiers were pseudonymised, not fully anonymous. Andrew describes NHS IT auditing CLIMB, two-factor authentication, restricted access and contractual commitments not to identify patients. Clinical laboratories retained the ability to link identifiers back to their records.
Reporting outbreaks, not just SNP counts
The reporting problem changed during the pandemic. Early conversations concerned lineages, genetic distances and possible transmission links; later, attention shifted towards variants of concern such as Alpha and Beta. Reporting that samples differed by two or three SNPs did not directly answer a clinician’s question about whether one patient infected another.
Andrew initially provided regional overviews of clusters and hotspots. A care-home investigation became a more concrete example: a lineage with two distinctive SNPs, combined with postcode mapping and patients’ ages, linked six care homes. The discussion identifies staff moving between sites as an explanation for spread despite restrictions on visitors.
It took weeks to engage the relevant public health body, by which time the outbreaks had finished. However, the investigation established contacts for faster responses later. The panel also considers the practical constraints on prevention, including staffing, sick pay and the subsequent prioritisation of care-home testing and sequencing.
Coordination across borders and vocabularies
Lee describes CDC’s role as providing information, technical support and coordination, while state health laboratories held authority for much of the response. Laboratories initially sent samples to CDC, then increasingly sequenced locally and deposited sequences in GISAID or NCBI. Torsten contrasts this with Australia’s working groups and committees, noting that Australia had no official national CDC at the time of recording.
The UK discussion distinguishes COG-UK’s coordinating work from the responsibilities of the four nations’ health agencies. A host observes that pathogens cross jurisdictions more readily than data: sharing agreements, material transfer agreements and other administrative arrangements were not always ready. As health agencies developed sequencing capacity, work moved back from academic partners.
The episode ends with another kind of translation problem: British geographical terminology, place-name pronunciation and regional differences in saying words such as data, phage and router.
Highlights
- 00:00:46 — The first recording with all three hosts together in person introduces the research-to-public-health theme.
- 00:02:15 — Travel histories and genomic evidence support investigations of transmission during flights.
- 00:05:06 — Torsten recalls an intensive laboratory course and his destruction award.
- 00:09:52 — Public health controls matter when genomic findings can contribute to lockdown decisions.
- 00:11:56 — Lee recalls investigating reinfection and finding a relevant preprint involving Torsten.
- 00:15:24 — Andrew explains how research networks and local hospital links brought his laboratory into COG-UK.
- 00:19:51 — CLIMB security auditing, two-factor authentication and safeguards for pseudonymised data.
- 00:21:17 — Communicating genomic relatedness evolves from lineage comparisons to variants of concern.
- 00:22:25 — Early regional reporting leads to an investigation linking six care homes.
- 00:26:06 — Lee describes CDC’s technical support and coordination with state health laboratories.
- 00:29:50 — Jurisdictional responsibilities and missing data-sharing arrangements complicate pandemic responses.
- 00:31:59 — British geographical terminology leads into a discussion of place names and pronunciation.
In their own words
Yeah, I did all three and so it made it clear that bioinformatics was the right career choice for me.
— Torsten Seemann, 00:06:28
Who is talking
- Lee Katz (host)
- Andrew Page (host)
- Nabil-Fareed Alikhan (host)
- Torsten Seemann (guest)
Also mentioned: Josh Quick, Nick Loman, Sharon Peacock.
Tools and resources mentioned
MinION, PCR, PFGE, capillary sequencing, ARTIC protocol, protocols.io, CLIMB, GISAID, NCBI.
Questions this episode answers
Why were consensus genomes not enough to investigate SARS-CoV-2 reinfection?
Torsten explains that a minor allele or mixed infection in an initial sample could be missed by its consensus sequence and become dominant later. Revisiting raw FASTQ data could help investigate this, but reliable metadata linking samples to the same patient was also needed.
How did Andrew Page’s research laboratory join the COG-UK response?
Existing research networks connected the team to COG-UK, while local hospital links and a nearby diagnostic unit provided access to samples. Andrew recalls reaching production sequencing within about eight days of the go-ahead and continuing weekly sequencing from April 2020 until mid-April 2022.
How was patient-linked sequencing data protected?
The research team generally received laboratory identifiers and limited metadata rather than directly identifying patient information. Andrew describes NHS IT auditing CLIMB, two-factor authentication and contracts governing access, while clinical laboratories retained the records needed to identify patients.
What did genomics reveal about transmission between care homes?
Andrew describes a lineage with two distinctive SNPs that, alongside postcode mapping and age information, linked six care homes. The panel discusses staff working across sites as an explanation for transmission despite restrictions on visitors, and notes that delays in engaging public health limited the immediate response.
What role did CDC play in the sequencing response described by Lee Katz?
Lee describes a coordinating and technical-support role, with state health laboratories responsible for much of the response. Samples were initially sent to CDC for sequencing, while laboratories later developed local capacity and deposited sequences in GISAID or NCBI.
This page was written from a machine transcript of the episode and checked against it. Listen to the episode for the whole conversation.
Show notes
For the first time ever all 3 MicroBinfies are together in person to record an episode. We are joined by Torsten Seemann for a conversation about how what we do in research can get lost in translation when applied to public health. We discuss what we did with SARS-CoV-2 genomics and somehow end up chatting about geography and language. Hope you enjoy.