Variant prevalence
Testing coverage and sequencing ought to be thought of when deciphering variant knowledge. Details about surveillance techniques for England is reported by the UK Well being Safety Company (UKHSA) within the National Influenza and COVID-19 surveillance report.
The prevalence of lineages amongst UK sequences by Phylogenetic Project of Named International Outbreak Lineages (Pangolin) designation is offered in Determine 1. Lineages are proven if there are greater than or equal to five,000 sequences since 6 March 2023 or in the event that they signify greater than or equal to 1% of sequences inside a single week during the last 6 weeks. Lineages that don’t meet these standards are mixed with their guardian lineage (for instance, BA.2.4 is mixed with BA.2).
The lineages have been assigned utilizing the correct Ultrafast Pattern placement on Present tRee (UShER) mode and model 1.22 of the Pangolin knowledge.
Determine 2 reveals the connection between all lineages proven in Determine 1. The determine particulars the hierarchical relationship of the lineages from left to proper, beginning with B.1.1 by way of 14 sublineage ranges, together with aliases and the guardian lineages for any recombinants included within the prevalence figures (for instance, BJ.1 and BM.1.1.1 are the guardian lineages for the XBB recombinant). The proportion of every lineage in the newest week of knowledge in Determine 1 is proven subsequent to the related lineage labels in Determine 2. The place a sublineage has a share in Determine 2, it’s excluded from the share given for any guardian lineages. Any lineages that aren’t current in the newest week of genomic knowledge won’t have a share in Determine 2. The colors in Determine 2 correspond to the lineage colors in Determine 1.
Determine 1. Prevalence of Pangolin lineages within the UK sequence knowledge with a specimen date from week starting 6 March 2023 to week starting 14 August 2023, as of 24 August 2023
The entire variety of legitimate sequence outcomes per week is proven by the black line. The ‘Different’ class on this plot incorporates all lineages that don’t meet the related standards after combining smaller sub-lineages. ‘Unassigned’ are sequences that would not be assigned a lineage by Pangolin. Lineages current in at the least 2% of sequences in the newest week are labelled to the precise of the plot.
Determine 2. Sankey diagram displaying the connection between Pangolin lineages in UK sequence knowledge with a specimen date between 14 August 2023 and 20 August 2023. Knowledge proven as of 24 August 2023.
Lineage colors match these in Determine 1.
V-23AUG-01 (BA.2.86)
BA.2.86 was first raised as a sign in monitoring on 14 August 2023 as a part of horizon scanning as a result of massive variety of mutations current within the out there worldwide sequences. It was subsequently raised from a sign in monitoring to a variant V-23AUG-01 on 18 August 2023 because of additional reported instances internationally and one case within the UK. Declaring this lineage as a variant facilitates detailed characterisation and evaluation.
Within the week starting 28 August 2023, 2 UK sequences have been categorized as BA.2.86 (knowledge as of 29 August 2023).
The lineage BA.2.86 doesn’t seem in Determine 1 or Determine 2 because it doesn’t meet the criterion of greater than or equal to five,000 sequences since 6 March 2023 or signify greater than or equal to 1% of sequences inside a single week during the last 6 weeks. Lineages that don’t meet these standards are mixed with their guardian lineage and due to this fact the UK BA.2.86 sequence is included within the BA.2 lineage counts.
A full variant definition for V-23AUG-01 is on the market on GitHub.
Variant modelling
Two fashions are at the moment used to estimate the expansion benefit of rising lineages: a logistic regression generalised linear mannequin (GLM) and a generalised additive mannequin (GAM). Fashions are match to a geographically stratified pattern of Pillar 1 instances to make sure that relative progress charges are estimated in relation to an area set of co-circulating lineages. Assessments related to journey are excluded. A full description of strategies might be discovered within the variant technical briefing series.
In latest months there are fewer exams to mannequin than earlier within the yr (see Determine 1). This is because of lowered sampling effort and decrease prevalence. Furthermore, the proliferation of lineages to watch signifies that pattern sizes for particular lineages might be small. Uncertainty in our modelled relative progress charges is due to this fact elevated, which is mirrored in bigger confidence intervals on the estimates.
We goal to pick lineages and/or teams of lineages which might be each particular sufficient to choose up on rising indicators however broad sufficient to maximise statistical energy. Any lineage that has made up greater than 1.5% of the full samples and with at the least 50 sequenced instances inside 6 weeks of the newest specimen date is modelled individually. Lineages that don’t meet these standards will probably be added to a bunch with their closest guardian lineage however won’t be aggregated additional than one. If the situation can’t be met for a selected lineage (for instance, it doesn’t have a detailed guardian lineage at excessive sufficient prevalence) it won’t be modelled.
Lineages with a special high-level guardian won’t ever be aggregated collectively (for instance, we won’t combination BA.2 and BA.5 to B.1.1.529). Unassigned lineages are excluded from this evaluation. Observe that these aggregations will usually be broader than within the prevalence plot offered in Determine 1. The relative progress price of broad lineage lessons (for instance, guardian lineages that embody many distant youngster lineages) will probably be much less informative than specific modelling of particular sub-lineages. This is the reason we set a restrict of 1 stage when aggregating youngster lineages into dad and mom. Strategies of lineage collapsing for the expansion price evaluation are due to this fact nonetheless being refined.
Progress charges have been based mostly on sequences sampled by way of Pillar 1 testing (primarily optimistic exams carried out in hospital) in England (Desk 1). The sampling vary for each the logistic regression GLM and GAM is from 4 March 2022 to 19 August 2023. The mannequin match for any lineage with a optimistic progress price benefit (with 95% confidence intervals (CIs) that don’t cross zero) are proven in Determine 3. The lineages with a optimistic progress price estimated with cheap confidence are XBB.1.16.6 (36.1%, GAM), EG.5.1 (35.36%, GAM), XBB.2.3.11 (32.03%, GAM) and EG.5.1.3 (26.86%, GAM).
Desk 1. Progress price (GR) of English sequence lineages as of 19 August 2023†
| Lineage* | Lineage group composition** | Pillar 1 pattern measurement*** | Weekly progress price benefit (GAM) | Estimated prevalence¥ (GAM) | Weekly progress price benefit (GLM) |
|---|---|---|---|---|---|
| XBB.1.16.6 | XBB.1.16.6 (100%) | 52 | 36.1% (95% CI: 31.9 to 40.29) | 5.48% (95% CI: 3.74 to 7.96) | 23.13% (95% CI: -17.95 to 64.22) |
| EG.5.1 (XBB.1.9.2.5.1) | EG.5.1 (64.38%); EG.5.1.4 (24.66%); EG.5.1.6 (8.22%); EG.5.1.2 (2.74%) | 146 | 35.36% (95% CI: 33.28 to 37.44) | 15.57% (95% CI: 12.37 to 19.42) | 36.37% (95% CI: 12.96 to 59.79) |
| XBB.2.3.11 | XBB.2.3.11 (90%); GS.1 (10%) | 60 | 32.03% (95% CI: 17.08 to 46.97) | 6.1% (95% CI: 3.03 to 11.89) | 49.32% (95% CI: 12.79 to 85.85) |
| EG.5.1.3 (XBB.1.9.2.5.1.3) | EG.5.1.3 (100%) | 58 | 26.86% (95% CI: 16.69 to 37.02) | 5.15% (95% CI: 3.15 to eight.33) | 22.61% (95% CI: -16.71 to 61.92) |
| FL.1.5 (XBB.1.9.1.1.5) | FL.1.5.1 (81.48%); FL.1.5 (18.52%) | 54 | 6.51% (95% CI: -9.87 to 22.9) | 3.02% (95% CI: 1.35 to six.62) | 7.39% (95% CI: -31.49 to 46.28) |
| GE.1 (XBB.2.3.10.1) | GE.1 (100%) | 113 | 4.17% (95% CI: -5.3 to 13.63) | 7.8% (95% CI: 5.39 to 11.16) | 8.98% (95% CI: -17.02 to 34.99) |
| XBB.1.16 | XBB.1.16 (65.53%); XBB.1.16.15 (8.9%); XBB.1.16.21 (6.25%); XBB.1.16.19 (5.11%); XBB.1.16.9 (3.79%)… | 528 | -1.92% (95% CI: -11.3 to 7.47) | 23.37% (95% CI: 17.12 to 31.05) | 1.41% (95% CI: -13.84 to 16.66) |
| EG.5.1.1 (XBB.1.9.2.5.1.1) | EG.5.1.1 (98.62%); HK.3 (0.92%); HK.2 (0.46%) | 218 | -14.06% (95% CI: -28.36 to 0.24) | 10.97% (95% CI: 5.8 to 19.76) | -13.85% (95% CI: -33.71 to six.01) |
| XBB.1.9.2 | XBB.1.9.2 (47.62%); EG.1 (21.43%); EG.2 (20.24%); EG.7 (8.33%); EG.5 (2.38%) | 84 | -17.55% (95% CI: -37.43 to 2.34) | 1.92% (95% CI: 0.92 to three.98) | -10.96% (95% CI: -48.24 to 26.33) |
| XBB.1.16.1 | XBB.1.16.1 (75%); FU.1 (18.75%); FU.2 (3.75%); FU.3 (1.25%); FU.4 (1.25%) | 80 | -21.91% (95% CI: -36.48 to -7.35) | 1.94% (95% CI: 1.04 to three.61) | -26.57% (95% CI: -63.77 to 10.64) |
| XBB.1.5 | XBB.1.5 (37.5%); XBB.1.5.28 (9.62%); XBB.1.5.11 (6.73%); XBB.1.5.71 (6.73%); XBB.1.5.59 (6.73%)… | 104 | -22.35% (95% CI: -33.2 to -11.5) | 2.55% (95% CI: 1.6 to 4.04) | -20.77% (95% CI: -52.27 to 10.73) |
| XBB.1.16.11 | XBB.1.16.11 (100%) | 62 | -70.7% (95% CI: -105.74 to -35.66) | 1.24% (95% CI: 0.23 to six.38) | 32.5% (95% CI: -1.62 to 66.63) |
*Listed guardian lineages embody all sub-lineages, apart from these explicitly modelled.
** The highest 5 contributing lineages to the modelled group in the newest 6 weeks (8 July 2022 to 19 August 2023). Greater than 5 sub-lineages are indicated by “…”
*** Pattern measurement is for Pillar 1 samples in England in the newest 6 weeks (8 July 2022 to 19 August 2023).
¥ Estimated prevalence for 19 August 2023.
† Sampling vary for each logistic regression and generalised additive fashions (GAM) is from 04 March 2022 to 19 August 2023.
CI = confidence intervals
Determine 3. Modelled prevalence of lineage teams with a progress price benefit over different circulating lineages
The black line reveals the central estimate and blue shaded areas the 95% confidence intervals. Factors present the nationwide stage proportions, with the scale being indicative of the pattern measurement for that exact lineage. The gray portion of the ribbon denotes that this time period is prone to be backfilled with extra sequenced instances, making proportions unreliable.
Sources and acknowledgements
Knowledge sources
Knowledge used on this investigation is derived from the Cloud Infrastructure for Microbial Bioinformatics (CLIMB) knowledge set, the UKHSA genomic programme knowledge set and the UKHSA Second Technology Surveillance System.
Authors of this report
UKHSA Genomics Public Well being Evaluation Workforce
UKHSA Infectious Illness Modelling Workforce
UKHSA TARZET Technical Secretariat
