German BA.2/BA.1/BA.2 recombinant with S:K147E, S:R346K, S:460K, S:493 reversion, S:H1101Y [55 seq as of 2022-07-22]

[JosetteSchoenma]

@FedeGueli spotted this one and mentioned it in #773
Description:
Recombinant between: BA.2/BA.1/BA.2
Earliest sequence: 1st of June 2022
Most recent sequence: 23th of June 2022
Country: Germany
Likely breakpoint: between 13195 and 15240 and between 21618 and 21762 [S:27 to S:67]
Private Mutations: C4927T, T5386G (like BA.1), C7834T, C12049T, A220001G/S:K147E, G22599A/S:R346K, G22992/S:S477N, C24863T/S:H1101Y, A27507G
Cov-Spectrum query: https://cov-spectrum.org/explore/World/AllSamples/Past3M/variants?nucMutations=C4321T%2CT5386G%2CC9344T%2CA18163G%2CG21987A%2CC24863T%2CA27507G%2CC12880T&variantQuery1=Nextcladepangolineage%3ABA.5&
Genomes:
EPI_ISL_13569221
EPI_ISL_13385893
EPI_ISL_13344896
EPI_ISL_13569385
EPI_ISL_13567682
EPI_ISL_13387015
EPI_ISL_13570913
EPI_ISL_13393217
EPI_ISL_13269776
EPI_ISL_13382835
EPI_ISL_13382506
EPI_ISL_13382561
EPI_ISL_13380484
EPI_ISL_13269235
EPI_ISL_13387937
EPI_ISL_13269123
Evidence:
Nextclade output:
Red is BA.1

Sc2rf [@corneliusroemer replaced Delta/Omicron sc2rf with relevant BA.1/BA.2]:

Usher Tree:
https://nextstrain.org/fetch/genome.ucsc.edu/trash/ct/subtreeAuspice1_genome_33c33_216ac0.json?branchLabel=Spike%20mutations&label=nuc%20mutations:C15240T

Original discussion in issue #773 with @FedeGueli and @silcn

[FedeGueli]

Thx for analysing these sequences! great job.
edited:
i would add that this sublineage has the S:H69- S:V70- 69/70del
and the S:493Q reverted to wildtype
both important features.
I am still not completely convinced that it is a simple recombinant event or something popping up from or next to the "Omicrons" source

@corneliusroemer @chrisruis @AngieHinrichs @rambaut @thomaspeacock

[silcn]

@FedeGueli apart from the fact this has only been found in Germany, another argument against a link to the Omicron source is the presence of 9866T in the BA.2 section. A combined BA.1/BA.2 chronic infection could certainly produce something like this.

[FedeGueli]

Good point thx @silcn .
@JosetteSchoenma from BA.1 there is also the 5386G that was highlighted by @c19850727 in the other issue.

[shay671]

In Israel we got a sample (will soon be uploaded to GISAID) of a sample clustering with this branch of a passenger coming from Greece (no contact to Germany)
(@FedeGueli ,@silcn@JosetteSchoenma)

[shay671]

If this is a real recombinant, it means that it's a recombinant + saltation. Because then we consider the 69/70 del not as part of BA.1 section (As the C21762T mutation of BA.1 is missing). And also it has many nonBA.2/BA.1 mutations which are highly converged in Omicrons 2nd gen/stepwise :
S:K147E
S:R346K
S:N460K
and the highly converged reversion in 493

[FedeGueli]

Thx @shay671 to me it is still contested if it is a plain recombinant , a recombinant in a chronic case or something coming from the entourage of original Omicron source.
The @silcn argument on Germany is weakened by the fact that BA.5+S:346I was barely sequenced in South Africa and then popped up and emerged fastly in Germany, represwnting a brand new branch of BA.5 distinct from the main designated or even proposed ones

More:
And i was just reading casually this paper yesterday :

"Potential evidence of a single recombination event involving BA.1, BA.2 and BA3 was identified by 3SEQ (P = 0.03), GARD (delta c-AIC = 20) and RDP5 (GENECONV P = 0.027; RDP P = 0.006) within the NTD encoding region of spike. The most likely breakpoint locations for this recombination event were 21690 for the 5′ breakpoint (high likelihood interval between 15716 and 21761) and 22198 for the 3′ breakpoint (high likelihood interval between 22197 and 22774)." (Tulio de Oliveira et Al:
https://www.nature.com/articles/s41586-022-04411-y)
the original BA.1/BA.2/BA.3 breakpoint seems to overlap with the breakpoints of this one described here. Sincerely i dunno know if significant or not, very far from me i would say to say anything relevant, but maybe someone else could take a 2nd look at this.

[silcn]

@FedeGueli BA.5+S:346I doesn't come directly from the Omicron source though. There have been at least 5 (maybe more, see #455, #561) separate emergences from the Omicron source and in every case most of the early sequences came from South Africa.

[FedeGueli]

@silcn yes true i mean if a very trasmissible lineage is barely detected in the country where it likely emerged i think a much lower trasmissible one can fly under the radar easily. Not saying it is wrong what have you said but that is not the strongest argument vs BA.6 hypothesis, i think. Alsoas you i think a new recomb in a chronic host different from Omicron source is the most likely explanation.

[JosetteSchoenma]

Thanks for all the comments! It gave me some extra insight.
I think it is more complicated than a normal recombinant with 2 breaking points. I missed the 69/70 del, since it is not in Sc2rf, but that would make it 2 BA.1 mutation clusters instead of one.
Together with the T5386G BA.1 mutation, which is shown by Sc2rf, and I did notice, but didn't highlight as BA.1 (just edited it in the first part of the issue) and the 493 reversion (A23040A) that really makes it very complicated.

To summarize the nucleotides:
1-4321 BA.2
T5386G BA.1
8393-13195 BA.2
15240-21618 BA.1
C21762C BA.2
21765-21770 BA.1 (69/70del)
21846-22882 BA.2
A23040A S:493Q reversion, no BA.1/2
23048-29510 BA.2
Plus of course several private mutations.

[silcn]

@JosetteSchoenma also possible that 69/70del occurred independently rather than being BA.1-derived. Wouldn't be too surprising given the other additional spike mutations.

[FedeGueli]

Thanks vm @JosetteSchoenma ! I suggest to edit your main comment so it would be clearer for everyone !

[FedeGueli]

@silcn yes that is possible, but in a landscape of multiple breakpoints and a key reversion for other Omicron lineages i dont know how much we can be sure of that or not.

[silcn]

@JosetteSchoenma also interesting that you mention BA.3, because something very similar happened there: it contains S:G446S from BA.1 in an otherwise ancestral BA.2-derived section. Again, impossible to know if it arose independently or though recombination. Still doesn't change my opinion that this lineage probably has nothing to do with the Omicron source (the C9866T is convincing enough for me).

The S:R493Q reversion doesn't make things any more complicated than they already are: it is extremely common in saltation-derived BA.2 sequences/lineages with multiple Spike mutations.

[shay671]

As I mentioned yesterday, a first Israeli sample (traveler from Greece) :

https://nextstrain.org/fetch/genome.ucsc.edu/trash/ct/subtreeAuspice1_genome_23033_861f50.json?c=userOrOld&label=nuc%20mutations:T15714C,T17410C,T21618C,A22001G,G22599A,C24863T,A27507G

[UlrichElling]

We have two cases in Austria as well, currently uploading to Github.

[FedeGueli]

@UlrichElling thx a lot!

@chrisruis @corneliusroemer @AngieHinrichs @InfrPopGen in my view this has to be granted an early designation, whatever as recombinant or something else, still very different from everything else circulating (or better saying that it has a unique mix of mutations actually on the rise) .

[UlrichElling]

What I find remarkable about this variant is that it appears to be convergent evolution with BA.2.75 (147, 460, 493) and with regards to the 493 reversion even BA.5. It seems obvious the virus has identified the next surfaces to optimise.

[shay671]

@UlrichElling
This is why it might be a combination of recombination+saltation in a chronic patient.

The mutations converged in this one are indeed more related to variants that arose outside SA and the parallel BA.1-5.

And the recombination is much cleaner than could be seen in cases where presumably a wide range of variants been co-evolved like might be in the original chronic patient ( like BA.3 probably coming from the same body where BA.1 and BA.2 evolved and XB probably coming from the same body were B.1.631 and B.1.634 evolved).

[shay671]

If anyone needs -here is a comparison of the mutational profile of this variant , BA.1 and BA.2
PDI 823 .xlsx

[FedeGueli]

Thx @shay671 !

[shay671]

For CovSpectrum query i would replace to :
[4-of:C4927T,C7834T,C12049T,A22001G,G22599A,T22942G,C24863T,A27507G]
searching 4 out of the 8 unique mutations in this lineage (not present in BA.1 or BA.2).

[JosetteSchoenma]

Here's a link for @shay671 's query.
https://cov-spectrum.org/explore/World/AllSamples/Past6M/variants?variantQuery=%5B4-of%3AC4927T%2CC7834T%2CC12049T%2CA22001G%2CG22599A%2CT22942G%2CC24863T%2CA27507G%5D&

[UlrichElling]

We found another one. Also I tried to illustrate a bit what the mutations are like:

[UlrichElling]

[FedeGueli]

23 sequences as today
Early signs of growth advantage vs BA.5 limited to Germany:
https://cov-spectrum.org/explore/Germany/AllSamples/from=2022-05-11&to=2022-07-20/variants?variantQuery=nextcladePangoLineage%3ABA.5*&variantQuery1=%5B4-of%3AC4927T%2CC7834T%2CC12049T%2CA22001G%2CG22599A%2CT22942G%2CC24863T%2CA27507G%5D&analysisMode=CompareToBaseline&

@chrisruis @corneliusroemer

[shay671]

27 samples as of yesterday

https://nextstrain.org/fetch/genome.ucsc.edu/trash/ct/subtreeAuspice1_genome_3a8c3_7e2120.json?c=pango_lineage_usher&label=nuc%20mutations:C15240T,A22001G,G22599A,C24863T,A27507G

GISAID ref.csv

[UlrichElling]

With the search term Spike_K147E, Spike_N460K, Spike_G339D, Spike_H69del currently a total of 60 cases are found. We mostly sequence only spike and selected regions in Austria, so these samples are only found with a term that does not demand too many mutations beyond spike. Could we maybe adapt the search term to account for that? 2 samples were also full-genome sequenced already. Really looking like growth advantage. Plus the spread across Germany and beyond says it all...

[FedeGueli]

@UlrichElling i would add Spike_H1101Y to your query.

[corneliusroemer]

@FedeGueli I would recommend using advanced query with NextcladePangoLineage:BA.5* instead of just two mutations which do not immediately show which lineage you compare to. I've edited your growth advantage query

@JosetteSchoenma I think your sc2rf plot is confusing as it has Delta BA.2 as parents rather than BA.1 and BA.2. Can you explain why you did that and maybe change it to BA.1/BA.2 if this is a mistake? It's confusing because the title of the issue suggests this is a BA.2/BA.1/BA.2 recombinant.

Here's a quick summary of how I think about this recombinant

Spike is mostly BA.2* but with some notable changes:

• Missing: T19I, 24-26del, 27S not there because that part is from BA.1* parent
• Donated: 69-70del convergent with many VOCs
• Additional S:R346K, convergent with BA.1.1
• Additional S:K147E, S:N460K and reverted S:493 convergent with BA.2.75

[corneliusroemer]

I've added a BA.1/BA.2 sc2rf pic to the first post in this issue to keep things simple for those reading from top to bottom @JosetteSchoenma

I tried to find which branch the BA.2* donor is from but couldn't find anything interesting in Germany/Europe. For the BA.1 part, there are no private mutations so if anything one can go by exclusion principle.

[JosetteSchoenma]

Thanks, @corneliusroemer . Somehow I forgot to include it. Couldn't have done my analysis without it. Here is the one I made from the first 16 sequences. It won't be very different.
https://user-images.githubusercontent.com/94831035/180068034-2dc94420-2149-4b9b-9ded-0bdaa539863e.png [hidden by @corneliusroemer for brevity of conversation]

[corneliusroemer]

@chrisruis @InfrPopGen @thomasppeacock

There are now at least 55 sequences of this recombinant, 37 have been submitted since Tuesday!

I would strongly recommend quick designation as the criteria are satisified, the Spike profile is very interesting and there appears to be growth, too.

Here are recent EPI_ISLs of complete sequences (3 are Spike only) so you can designate even if these sequences aren't yet in the Usher tree (maybe excluded due to long branch? @AngieHinrichs

EPI_ISL_13269123, EPI_ISL_13269776, EPI_ISL_13344896,
EPI_ISL_13380484, EPI_ISL_13382506, EPI_ISL_13382561,
EPI_ISL_13382835, EPI_ISL_13385893, EPI_ISL_13387015,
EPI_ISL_13387937, EPI_ISL_13393217, EPI_ISL_13567682,
EPI_ISL_13569385, EPI_ISL_13570913, EPI_ISL_13795661,
EPI_ISL_13910672, EPI_ISL_13911687, EPI_ISL_13913622,
EPI_ISL_13917987, EPI_ISL_13919465, EPI_ISL_13923147,
EPI_ISL_13923171, EPI_ISL_13923416, EPI_ISL_13923424,
EPI_ISL_13923524, EPI_ISL_13923983, EPI_ISL_13924147,
EPI_ISL_13927733, EPI_ISL_13933571, EPI_ISL_13933603,
EPI_ISL_13933607, EPI_ISL_13933624, EPI_ISL_13933724,
EPI_ISL_13935440, EPI_ISL_13936383, EPI_ISL_13936388,
EPI_ISL_13936925, EPI_ISL_13937107, EPI_ISL_13937273,
EPI_ISL_13937536, EPI_ISL_13938846, EPI_ISL_13938960,
EPI_ISL_13939246, EPI_ISL_13939997, EPI_ISL_13941569,
EPI_ISL_13942052, EPI_ISL_13943854, EPI_ISL_13943861,
EPI_ISL_13944742, EPI_ISL_13945288, EPI_ISL_13952188-13952189,

[AngieHinrichs]

Thanks @corneliusroemer -- yep, they are filtered out due to having too many reversions relative to BA.2 placement by nextclade, but I've been manually exempting the EPI_ISL's from Josette's initial list and EPI_ISL_13719505 from the cov-spectrum query as of a few days ago. They're in the tree, on a branch with the label proposed823 (Pango lineage assigned by UShER).

My list has a few that are not in your list:
EPI_ISL_13269235
EPI_ISL_13569221
EPI_ISL_13719505

I will exempt the rest from your list, so they should be added in tomorrow's build.

[corneliusroemer]

Thanks @AngieHinrichs! My list was created using manual AA substitution queries in GISAID - not super reliable. So not surprised you catch a few more.

[FedeGueli]

The @shay671 query finds now 63 sequences of this lineage:
https://cov-spectrum.org/explore/World/AllSamples/Past6M/variants?variantQuery=%5B4-of%3AC4927T%2CC7834T%2CC12049T%2CA22001G%2CG22599A%2CT22942G%2CC24863T%2CA27507G%5D&

[InfrPopGen]

Thanks for submitting. We've added recombinant lineage XAK with 55 newly designated sequences, and 0 updated designations