Skip to Main Content

SARS-CoV-2: Variant Soup

Information and resources on SARS-CoV-2 (the virus that causes COVID-19)


For the most up-to-date information on the global trends of COVID-19 variants and mutations, check Cornelius Roemer's monthly SARS-CoV-2 Variant Reports on GitHub.

Latest Update: August 23, 2023


BA.2.86 Update
Sequencing Count Data

Below is a table summarizing BA.2.86 count data compared to all sequences shared via GISAID per calendar week of collection by geography region.

Date of analysis: 2023-08-23, 18:00 UTC, GISAID data was queried via the GISAID web interface.

Week of: Global Europe N. America S. America Asia Africa Oceania Denmark S. Africa
2023-07-10 0 / 9,671 0 / 1,543 0 / 3,523 0 / 141 0 / 3,942 0 / 21 0 / 479 0 / 44 0 / 13
2023-07-17 0 / 8,727 0 / 1,543 0 / 3,435 0 / 58 0 / 3,373 0 / 21 0 / 297 0 / 23 0 / 16
2023-07-24 3 / 8,232 1 / 1,606 1 / 2,879 0 / 29 0 / 2,691 2 / 12 0 / 261 1 / 49 2 / 11
2023-07-31 3 / 5,942 1 / 1,411 1 / 2,879 0 / 14 1 / 1,386 0 / 8 0 / 170 1 / 45 0 / 5
2023-08-07 3 / 2,756 2 / 480 1 / 1,607 0 / 22 0 / 553 0 / 11 0 / 83 1 / 16 0 / 6
2023-08-14 0 / 306 0 / 65 0 / 184 0 / 2 0 / 50 0 / 0 0 / 4 0 / 0 0 / 0
Phylogenetic Analysis

You can find a Nextstrain tree of all 9 currently available consensus sequences shared via GISAID at A screenshot is shown below:

The time of most recent common ancestor can be estimated from the number of mutations between the ancestor and the tips with known dates. All but one sequence are within 2-4 mutations from the common ancestor. Within variants, mutations can be modelled as arising through a Poisson process with a rate of about 17 mutations per year (see Neher, Contributions of adaptation and purifying selection to SARS-CoV-2 evolution). The common ancestor can hence be dated to early May to mid June.


The count data show that at the beginning of August, BA.2.86 was likely to account for between 1 in 10,000 and 1 in 1000 SARS-CoV-2 cases on the continents that have the densest genomic surveillance: North America and Europe (Asia sequences similar numbers of samples but has higher heterogeneity: the majority comes from only a few countries). When excluding the US-airport surveillance sequence, two countries have had more than 1 detection to date: Denmark and South Africa. Given the small numbers of sequences, it is currently not possible to draw firm conclusions about the potential transmission advantage of BA.2.86 over other variants in these countries.

SARS-CoV-2 Monthly Variant Report

Latest Update: 26 July 2023

This month's report provides an overview of the main XBB and non-XBB branches currently circulating, covering: XBB.1.5, XBB.1.9.1, XBB.1.9.2, XBB.1.16, XBB.1.22.1, XBB.2.3, XBB.1.18.1, XBB.1.17.1, XBB.1.42, XBB.1.19; as well as non-XBB lineages CH.1.1, BN.1, BA.2.3.20, XAY and XBC.

In addition to the branches themselves, this report also examines beneficial spike mutations occurring repeatedly in various XBB. The mutations discussed include S:478R, S:F456L, S:E554K, S:Q613H, S:S704L, S:A701V, S:K356T, and S:K403R.

Main XBB Branches
XBB.1.5 (Nextstrain clade 23A, S:486P S:252V)

XBB.1.5 was the first XBB sublineage to acquire S:F486P and spread widely. It was most common in USA/Canada, where it peaked at around 80% in March 2023. The global unweighted average has decreased from a peak of ~55% in March to ~20% at the end of June.

Interesting sublineages:

  • XBB.1.5.70 (Pango alias GK) with S:L455F, S:F456L, most common in Brazil, where it has grown to around 30%.
XBB.1.9.1 (Pango alias FL, 486P 252V)

Shares ORF9b:I5T with XBB.1.9.2 and XBB.1.16, and ORF1a:G1819S + ORF1a:T4175I with XBB.1.9.2. Peaked globally at around ~15-20% in May/June. Historically most common in South East Asia (Indonesia/Malaysia).

Interesting sublineages:

  • FL.1.5.1 (Pango alias HN) with S:F456L, S:478R, S:A701V, most common in Dominican Republic where it appears to have become dominant. On the rise in US/Canada.
XBB.1.9.2 (Pango alias EG, 486P 252V)

Shares ORF9b:I5T with XBB.1.9.2 and XBB.1.16, and ORF1a:G1819S + ORF1a:T4175I with XBB.1.9.2 (acquired 486P independently, but differs from XBB.1.9.1 otherwise only by synonymous mutations). Globally on the rise with 15-20% share at the end of June. Historically most common in Indonesia, currently most common in China (in particular sublineage EG.5.1).

Interesting sublineages:

  • EG.5.1 with S:F456L, S:Q52H, ~30% share in China, ~10% globally at end of June
XBB.1.16 (Nextstrain clade 23B, 478R 486P 180V 252V)

Shares ORF9b:I5T with XBB.1.9.1/XBB.1.9.2. Historically most common in India where it reached ~80% in April. First XBB lineage with both 486P and 478R to spread widely. Globally stable at around 20-25% at the end of June.

Interesting sublineages:

  • XBB.1.16.6 with additional S:F456L (via T22930A), in the US already ~4% at end of June
  • XBB.1.16.9 with additional S:F456L (via T22930G), most common in Singapore, at around 2% at end of June
XBB.1.22.1 (Pango alias FY, 486P 200C 252V)

Shares ORF9b:I5T with XBB.1.9.1/XBB.1.9.2/XBB.1.16. Most common in Asia, particularly China, where it has grown to above 10% at the end of June.

Interesting sublineages:

  • FY.4.1 (S:Y451H, S:494P), dominant in Kenya in March/April
XBB.2.3 (486P 521S 253G)

Only common XBB sublineage that has 252V instead of 253G. Only main XBB lineage to have ORF1b:P959S reverted to Wuhan (mutation inherited from the BJ.1 parent, where it was the defining mutation of BA.2.10). Most common in India where it has been at around 20% since January. Globally on the rise with around 10% share at the end of June.

Interesting sublineages:
  • XBB.2.3.8 with S:478R
  • XBB.2.3.11 with S:478R
  • GE.1 (XBB. with S:478R and S:185-, S:F186I due to 3nt cross-codon deletion
XBB.1.18.1 (Pango alias FE, S:486P 252V)

Independently acquired 486P after ORF1a:D2579G. Most common in Brazil where it reached >50% share in May. Globally just below 2% at the end of June, though likely higher if weighted by population rather than raw sequencing volume due to Brazil sequencing less than the global average (7 seqs from past 3 months per million inhabitants vs 18 globally).

Interesting sublineages:

  • FE.1 with S:F456L, dominant in Brazil, having effectively replaced all other XBB.1.18.1, ~2% globally
Other XBB sublineages
XBB.1.42 (S:486P 613H)

Shares ORF9b:I5T, ORF1a:T4175I with XBB.1.9.1/2 but lacks ORF1a:G1819S. Has known slightly beneficial mutation S:Q613H that has consistently conferred a slight growth advantage to various variants over the past 2 years.

Most common in China at nearly 3% at end of June, but also present in Japan, US and others at >1%. Globally at around 1% at end of June.

Interesting sublineages:

  • XBB.1.42.1 (S:478R)

XBB.1.19 (S:486P 252V)

Only pre-ORF8-stop XBB.1 lineage on this list. Most common in Pakistan where it appeared to be at 40-50% in April. Majority is sublineage XBB.1.19.1 (GW) with additional S:E554K (mutation known to be slightly beneficial mutation). Globally at ~0.5% (unweighted average) but this is certainly an underestimate given Pakistan represents more than 2% of global population.

Interesting sublineages:

  • GW.5 (XBB. with S:478I, S:L455F, S:F456L

XBB.1.17.1 (Pango alias GA, S:486P 215H)

Appears to be dominant in Central/West Africa (Mali, Togo, Nigeria, Ghana, Burkina Faso, Cote d'Ivoire) at least between January and April 2023 (lack of data precludes more recent analysis). Globally at around 0.5% at end of June but due to low sequencing volume where the lineage is most common the real population weighted share is more likely to be between 1-5%.

Interesting sublineages:

  • GA.4 (S:K356T)
Non-XBB Branches

While XBB has been dominating global circulation for the past 6 months, there are still a number of non-XBB branches that are persisting, albeit at low levels. Interestingly, the global XBB-share seems to be no longer increasing and stable at around 95%.

CH.1.1 (BA.2.75 with 346T, 444T, 452R, 486S)

CH.1.1 reached its global peak in February 2023 at 8% when it was most common in the UK at around 30%. It declined from there to around 3% in April, a level it has been stable at since then. The actual population weighted global average is however likely lower as CH.1.1 is most common in New Zealand and Australia, two countries that sequence much more than the global average. The globally most common sublineage is currently FK.1.1 which is defined by extra Spike mutations S:D215G and S:Q613H. In New Zealand, in the most recent data from mid June, CH.1.1* was at around 35%. In Australia, CH.1.1* at around 15%.

Interesting sublineages:

  • FK.1.1 (S:D215G, S:Q613H), at stable level in New Zealand, Australia, competitive with XBB lineages.
  • DV.7.1 (S:N185D, S:L858I, S:F455L, S:F456L), growing in Spain, albeit at low level
  • GQ.1 (S:S255P, S:486P)
BN.1 (BA.2.75 with 346T, 356T, 490S)

BN.1* reached its global peak in January 2023 at 6%, it reached 60% in South Korea in February 2023 and 20% in Japan in March 2023. Recently, it has been most common in China and Taiwan at around 2% at the end of June 2023. Globally, it has slowly declined below 1%. The currently most interesting sublineage is FR.1, defined by an extra S:F187R. FR.1* reached around 4% in China at its peak in May 2023.


BA.2 descendant with M153T, N164K, H245N, G257D, K444R, N450D, L452M, N460K, A484R, R493Q. BA.2.3.20 was the dominant lineage in the Philippines until February 2023 when it was outcompeted by XBB. It has since declined to around 0.1-0.2% globally. While BA.2.3.20* overall has declined, some sublineages have evolved that appear to be competitive in the current variant landscape, at least at low levels. The most competitive sublineage is FV.1 (with extra S:K147N, S:R346T, S:G446S, S:F486S) which was first observed in the Philippines but has since established itself at low levels in the Dominican Republic (situated on the Caribbean island Hispaniola which it shares with Haiti) as well as Japan, South Korea and Hong Kong and appears to be growing.


XAY is a Delta-Omicron recombinant. The Spike NTD is from Delta (breakpoint between 159 and 212), the rest of the Spike is BA.4/5-like. XAY has important extra mutations in the Spike beyond what's derived from the donors: in particular S:F486P (known as XBB.1.5-defining), and S:G446D (XBB has similar S:G446S). XAY was first sequenced in South Africa in June 2022 where it made up around 1% of sequences in the second half of 2022, peaking at 2% around September 2022. At the end of 2023, XAY appeared in Europe, in particular Denmark, where XAY.2 reached a peak of 4% in January/February 2023. XAY.1.1 with extra S:D253G and S:R346T, established itself in Germany in February/March/April 2023 at around 1%, Most of it being XAY.1.1.1 with extra S:D1153Y. XAY* was only sporadically observed on other continents.

GL.1 (XAY. with extra S:D420N and S:Y144- was first sequenced in Spain in Alicante in March 2023. It has grown to around 0.3% in Europe by end of March and appears particularly common in South Europe (Portugal, Spain and Italy). It has been sporadically observed in North America and Australia.


XBC is a Delta(21I)-Omicron(BA.2) recombinant with 3 breakpoints. The Spike NTD is from 21I-Delta, while the rest of Spike starting between S:340 and S:370 is BA.2-like with extra Spike mutations S:G446S, S:486P, S:R493Q (reversion). It was first observed in the Philippines in samples collected in August 2022. In the Philippines, it was at around 10-20% in the Philippines in September to November 2022, stayed between 5-10% between December and March and further decreased to around 1% since. In contrast to the slow decrease in the Philippines, XBC has steadily grown in Australia: from around 1% in October 2022 to 10% in March 2023, growing further to above 30% by the end of June. The most competitive sublineage of XBC is XBC.1.6 with extra S:R346S and S:L452R (via XBC.1's S:L452M). While XBC.1.6 has grown steadily in Australia from 1% in February 2023 to 30% at the beginning of July 2023, it has only had limited success outside of Australia. XBC.1.6 only represents 0.1% of sequences with collection date since June 1 2023 in Europe, 0.2% in North America and 0.1% in Asia.