SARS-CoV-2 variants and genomic surveillance in Africa
Recorded on 19 February 2021, this episode brings Peter van Heusden together with hosts Lee Katz, Andrew Page and Nabil-Fareed Alikhan for an update on SARS-CoV-2 genomics. They discuss emerging variants, interpreting weak sequencing evidence, and the difficulty of sustaining genomic surveillance across Africa. The conversation distinguishes signals worth investigating from evidence strong enough to support public health conclusions.
Listen to Episode 48: SARS-CoV-2 More Variants of Concern and updates from Africa on SoundCloud
In this episode
Emerging variants and recurring mutations
The roundup begins with B.1.1.7 carrying the spike mutation E484K, nicknamed Bleak. One host describes a linked outbreak in Bristol alongside independent appearances elsewhere in the UK. E484K is also discussed in relation to B.1.351, P.1 and P.2, with concern about its implications for vaccine performance.
The discussion broadens to B.1.525, associated with Nigeria or travel there, A.23.1 in Uganda, and C.2 in Zimbabwe. Not all these lineages carry E484K. The panel asks when an expanding lineage should count as concerning: a superspreading event can produce rapid growth without demonstrating an intrinsic advantage for the virus.
Interpreting the South African vaccine trial
Peter describes the uncertainty surrounding the South African phase II AstraZeneca vaccine trial, in which he is a participant. He does not know whether he received vaccine or placebo. The trial involved about 2,000 people, with roughly 1,000 across both arms in the subgroup analysis following the emergence of B.1.351.
The reported efficacy estimate had very wide confidence limits, extending from around −40% to more than 50%. Peter stresses the small sample and the lack of statistical power to assess protection against severe disease. Participants were generally healthy, with an average age in their thirties. Against this uncertain evidence, he reports plans to donate South Africa’s AstraZeneca doses to the African Union.
Making cautious use of low-coverage data
One host describes searching marginal Illumina data for E484K, sometimes down to a single supporting read. Several lineage-defining mutations appearing together can provide a useful signal even when a sample fails normal quality criteria. Such findings are passed to public health as high risk, low confidence, rather than treated as ordinary database submissions.
Public Health England’s published thresholds distinguish different levels of evidence for identifying variants. The discussion considers situations where 30–40% of samples fail GISAID criteria but may still contain useful information. Initial inspection with mpileup developed into spreadsheets and reports that make interesting mutations easier to spot.
Peter distinguishes one-read coverage from a single read contradicted by other reads, and says 5× coverage would give him more confidence. Wastewater is suggested as a possible early-warning application, although degraded RNA and poorly performing amplicons complicate that work.
Names, bird lineages and recombination claims
The American birds preprint introduces nicknames for sub-lineages with spike changes at position 677. Peter connects this discussion to B.1.525, which also carries Q677H. The panel contrasts memorable names with Pango lineages, detailed UK lineage numbering and clade labels such as 20G and 20B. Geographic names become especially awkward when several distinct variants emerge in the same country.
Claims of recombination receive particular scrutiny. One host wants underlying reads, including linked mutations on a single high-quality read where possible, rather than an unusual consensus sequence alone. Cross-contamination, mixed infections and sequencing-run artefacts are alternative explanations that need checking before a striking sequence becomes a public announcement.
Sustaining surveillance across Africa
Peter credits continuing sequencing efforts at KEMRI in Kenya, in Ghana, and at UVRI in Uganda. Their value lies not only in producing initial genomes but in returning with further samples over time. Elsewhere, promising work has stalled because reagents ran out, shipping was difficult or research funds used for emergency surveillance were exhausted. Moving from an academic initiative to routine public health provision remains a central challenge.
B.1.351 is reported moving northwards into Botswana and Zambia; the resources include a Zambia sequence report in MMWR. Sequencing with Zimbabwe for December and January found it in about 60% of samples in December, rising to about 90% at the time of recording. A.23.1 was also observed, alongside a lineage under investigation carrying a 501T mutation. Peter contrasts this with Uganda, where A.23.1 predominated, and reports that neither B.1.351 nor B.1.1.7 had taken over in the Ugandan and Kenyan data being discussed.
Travel connections and surveillance gaps
Reports from Ghana describe B.1.1.7 in both community cases and travellers. Peter argues that variant distributions reflect human connections: flights link countries with the UK, Europe and Dubai, while spread from South Africa into neighbouring countries may follow road travel. Broader sequencing helps connect apparently isolated detections across countries.
Nabil-Fareed Alikhan reports that only 4.4% of the GISAID records he tabulated came from low- and middle-income countries. India ranked first and South Africa second among those countries by absolute sequence count, not population-adjusted contribution. The episode closes by noting both the large geographic blind spots and a Western and anglophone bias in how these conversations take place.
Highlights
- 00:00:47 — Recording date, the Africa focus and Peter van Heusden’s introduction
- 00:01:37 — B.1.1.7 with E484K and the Bleak nickname
- 00:02:58 — Uncertainty in the South African AstraZeneca trial results
- 00:06:19 — Searching low-coverage Illumina data for important mutations
- 00:07:24 — Public Health England thresholds for low-quality variant evidence
- 00:11:45 — B.1.525, Q677H and the American bird-named lineages
- 00:13:56 — Pango naming and competing lineage classification schemes
- 00:18:20 — Read-level checks and contamination concerns in recombination claims
- 00:21:00 — Continuing surveillance in Kenya, Ghana and Uganda
- 00:22:51 — B.1.351 dominance in Zimbabwe’s December and January samples
- 00:24:20 — Border checks, Ghana’s B.1.1.7 detections and travel connections
- 00:26:06 — Low- and middle-income countries’ 4.4% share of tabulated GISAID data
In their own words
And we really do need to have surveillance just absolutely everywhere.
— one of the hosts, 00:25:24
Who is talking
- Lee Katz (host)
- Andrew Page (host)
- Nabil-Fareed Alikhan (host)
- Peter van Heusden (guest, South African National Bioinformatics Institute)
Tools and resources mentioned
GISAID, Illumina, Nanopore, mpileup, Pango lineage nomenclature, UK lineage nomenclature.
Questions this episode answers
What did Bleak mean in this episode?
Bleak was a nickname for B.1.1.7 carrying the spike mutation E484K. The hosts discuss a Bristol outbreak and independent appearances elsewhere in the UK.
Can low-coverage SARS-CoV-2 sequences still inform public health?
The panel describes using several lineage-defining mutations in very low-coverage data to flag high-risk, low-confidence samples. This is presented as a targeted public health signal, not equivalent to a high-quality genome suitable for routine database submission.
What limited interpretation of the South African AstraZeneca trial?
Peter highlights the small subgroup analysis, wide confidence limits and generally healthy participant population. The study provided insufficient evidence to determine protection against severe disease caused by the emerging variant.
What made sustained genomic surveillance difficult in Africa?
Peter describes reagent shortages, shipping constraints and research funding that ran out after initial sequencing efforts. The challenge was to turn successful early work into continuing surveillance supported beyond individual research projects.
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
We discuss the latest developments in SARS-CoV-2 genomics over the last 2 weeks with Peter van Heusden, covering the growing list of Variants of Concern and the latest developments in Africa. Papers & resources mentioned: American birds (677) - https://www.medrxiv.org/content/10.1101/2021.02.12.21251658v2 https://github.com/cov-lineages/pango-designation PHE thresholds for different variants: https://www.gov.uk/government/publications/covid-19-variants- genomically-confirmed-case-numbers/variants-distribution-of-cases-data A.23.1 in Uganda - https://www.medrxiv.org/content/10.1101/2021.02.08.21251393v1 B.1.525
- https://github.com/cov-lineages/pango-designation/issues/4 Zambia sequences - https://www.cdc.gov/mmwr/volumes/70/wr/mm7008e2.htm?s_cid=mm7008e2_w