XBB.1.5 sublineage with ORF1a:A4068S, S:L455F, S:F456L, and ORF8:P38S (>70 seq. Chile, USA, South Korea, Brazil, Finland)

[oobb45729]

EPI_ISL_17589777
EPI_ISL_17535477
EPI_ISL_17500152
EPI_ISL_17593296
Substitutions on top of XBB.1.5: C27059T then ORF8:P38S then C1973T+C7564T+T8737C+ORF1a:A4068S(G12467T)+C22624T+S:L455F(G22927T)+S:F456L(T22928C)+A25045G
https://cov-spectrum.org/explore/World/AllSamples/Past6M/variants?variantQuery=ORF8%3AP38S%26ORF1a%3AA4068S%26S%3AL455F%26S%3AF456L&


https://nextstrain.org/fetch/genome.ucsc.edu/trash/ct/subtreeAuspice1_genome_41be3_3fd7b0.json?c=gt-S_455&label=id:node_6048398
2 sequences in 2 states of USA further acquired ORF1a:R287G, ORF1a:E633A, ORF1b:D1130G, S:S704L, ORF3a:T89I.

[FedeGueli]

Already Spotted by @krosa1910 in #1681

[krosa1910]

Thanks for referring me @FedeGueli
I proposed issue #1925 on 4/20 basically for L455F and F456L. At that time, the Chilean sequence was considered an out-group by me as it was placed in a different sub-tree, got a lot of new mutations and it was 1 seq at then. I decided to discard that, while #1925 ultimately coincide with #1681, and I closed it.
Then when I was searching yesterday, I found that the Chilean sequence had grown into 4 seqs with 3 US seqs. I questioned whether it evolved independently or was a continuation and asked Fede to verify.
If this is classified as a distant branch of issue 1681, then I think Fede should reopen his issue and treat this as duplicated, if it arises independently then you could keep this one. Anyway I want new potentially strong lineages to be noticed and assigned, and from Yunlong Cao's work this looked promising to me.
Also, one new seq in South Korea, also quite recent.

[oobb45729]

L455F(L455W) and F456L(F456V) are often seen together. In #1614 I talked about how L455F+F456L may be synergistic.

L455 and F456 are important for ACE2-binding. They fill a gap between RBD and ACE2. To optimize ACE2-binding, these two residues need to be: 1. hydrophobic. 2. not too large or too small. Take a look at how BA.2 RBD interacts with ACE2. (PDB: 7XB0)
F455+F456 might be too large while L455+L456 might be too small. However, L455F+F456L means that L455 and F456 swap their places, so they as a whole wouldn't be too large or too small.

I used AlphaFold to predicted how F455+L456 variants may interact with ACE2.

Comparing to L455+F456 variants, the double mutations do not disrupt the interaction too much.

[krosa1910]

L455F(L455W) and F456L(F456V) are often seen together. In #1614 I talked about how L455F+F456L may be synergistic.

L455 and F456 are important for ACE2-binding. They fill a gap between RBD and ACE2. To optimize ACE2-binding, these two residues need to be: 1. hydrophobic. 2. not too large or too small. Take a look at how BA.2 RBD interacts with ACE2. (PDB: 7XB0)
F455+F456 might be too large while L455+L456 might be too small. However, L455F+F456L means that L455 and F456 swap their places, so they as a whole wouldn't be too large or too small.

I used AlphaFold to predicted how F455+L456 variants may interact with ACE2. Comparing to L455+F456 variants, the double mutations do not disrupt the interaction too much.

This is very interesting to learn. I also thought that L/F interchange at two major escape neighboring position is beneficial, but don't have the tools to verify that.
Anyway, I though that probably the Pakistani 1681 failed because it lacked the star mutation F486P and thus probably lacked hACE binding to spread. This one, however got those from XBB.1.5 and I think it could be a thing.

[FedeGueli]

@oobb45729 that was true since the time of Delta. @c19850727 ( at that time i was not admitted to use Gisaid) found a bunch of late delta lineages with that double mutation (or 456V) and sometimes they showed up often with 446V too.

[krosa1910]

@oobb45729 that was true since the time of Delta. @c19850727 ( at that time i was not admitted to use Gisaid) found a bunch of late delta lineages with that double mutation (or 456V) and sometimes they showed up often with 446V too.

want to say something more about the nature of this weird "swap".
This is definitely not one of the very popular double mutation series. Rising only 105 times within the entire pandemic/over 15 million seqs. 30 of those are from this year out of half a million seqs.
However, only 16 of those are within the XBB family that seemed to dominate our world now. 11 are in family 1( some pre-486P XBB, #1681 ) and 5 in this nomination.
What else got those? 1 BN.1.3.X, 1 BA.1, 1 BA.4.6, 3 BA.5.2.1, 1 BA.5.1, 4 BQ.1.1* and 2 AY, yes Delta. Many of those are variants that have long gone to grave.
What's shared between those non XBB variants that got this combo? All but one got L452R, hallmark Delta mutation。
And these Delta are probably the most modern Delta I have ever seen. The earlier one even got S:477N, classic BA.2 move.

@FedeGueli Would you like to comment more about how this could be related to cryptics?

[krosa1910]

1 extra seqs submitted from Florida May-7

[ryhisner]

This appears to be dominant in the region of Ceara, Brazil—it makes up 51 of 73 sequences collected there in April. I think this definitely merits designation now.

[FedeGueli]

cc @corneliusroemer ping

[ryhisner]

Here's a wider, then closer up view of the current tree. Over 40% of the sequences (31) are in the upper branch, which has ORF3a:T89I and the ever-convergent S:S704L, and given the apparent rapid growth of this lineage in at least one region of Brazil, this branch seems likely to deserve its own designation before very long.

Only 11 of the Brazilian sequences have coverage at S:230, and three of those have S:P230T, so it's possible that this mutation is found in a high proportion of these sequences.

Permalink to the current Usher tree: https://nextstrain.org/fetchraw.githubusercontent.com/ryhisner/jsons/main/XBB.1.5_L445F_F456L.json

[oobb45729]

Maybe a closer look at those N417Ks is needed.

In NY3, the T417K reversion appeared along with the 455/456 double mutations.

[ryhisner]

Not impossible, but I'd be shocked if N417K turned out to be real. The same artifactual reversion shows up in sequences from every other lineage in Brazil. I checked the most recent several hundred sequences form this lab, and N417K turns up in every lineage they've sequenced in the past 6+ months.

[c19850727]

IMHO that subbranch with S:S704L and ORF3a:T89I maybe already deserve a proposal.

[FedeGueli]

IMHO that subbranch with S:S704L and ORF3a:T89I maybe already deserve a proposal.

@corneliusroemer @InfrPopGen it would be great to get the main branch designated

[FedeGueli]

Maybe a closer look at those N417Ks is needed. In NY3, the T417K reversion appeared along with the 455/456 double mutations.

was thinking the same . not sure unreal

[FedeGueli]

Gisaid query Spike_L455F,Spike_F456L ,NSP8_A126S finds 76 seqs

[FedeGueli]

@c19850727 found reports of oxygen shortages in Hospitals due flu like syndrome in a northern region of Brazil named AMAPA',while in the nearby state of Para' the lineage proposed here was sequenced the second time in Brazil back in March ( @ryhismer found this)
. While obviously the link between the two things is very weak the Upload doubling time for this lineage seemed to become shorter and shorter, @corneliusroemer for these reasons i suggest a rapid designation of this one. cc @thomasppeacock @InfrPopGen @AngieHinrichs

[InfrPopGen]

Thanks for submitting. We've added lineage XBB.1.5.70 with 5 newly designated sequences, and 0 updated. Defining mutations G12467T (ORF1a:A4068S), G22927T (S:L455F), T22928C (S:F456L), C1973T, C7564T, T8737C, C22624T, A25045G (following C27059T). Only 5 good qual sequences on the usher tree at the moment. If the pending sequences aren't correctly assigned later, I will designate some more.

[fiuzatayna]

Thanks for submitting. We've added lineage XBB.1.5.70 with 5 newly designated sequences, and 0 updated. Defining mutations G12467T (ORF1a:A4068S), G22927T (S:L455F), T22928C (S:F456L), C1973T, C7564T, T8737C, C22624T, A25045G (following C27059T). Only 5 good qual sequences on the usher tree at the moment. If the pending sequences aren't correctly assigned later, I will designate some more.

I see 70 sequences on GISAID with the mentioned mutations (G12467T,G22927T,T22928C,C1973T,C7564T,T8737C,C22624T,A25045G,,C27059T). I apologize if it is a simple question: shouldn't they be re-classified as well?

[FedeGueli]

Thanks for submitting. We've added lineage XBB.1.5.70 with 5 newly designated sequences, and 0 updated. Defining mutations G12467T (ORF1a:A4068S), G22927T (S:L455F), T22928C (S:F456L), C1973T, C7564T, T8737C, C22624T, A25045G (following C27059T). Only 5 good qual sequences on the usher tree at the moment. If the pending sequences aren't correctly assigned later, I will designate some more.

I see 70 sequences on GISAID with the mentioned mutations (G12467T,G22927T,T22928C,C1973T,C7564T,T8737C,C22624T,A25045G,,C27059T). I apologize if it is a simple question: shouldn't they be re-classified as well?

It will take some time, maybe weeks until Gisaid updates to the last pango release and there should have been a new pango release before that. But you will see soon on Nextclade and CovSpectrum and from tomorrow on Usher

[InfrPopGen]

I see 70 sequences on GISAID with the mentioned mutations (G12467T,G22927T,T22928C,C1973T,C7564T,T8737C,C22624T,A25045G,,C27059T). I apologize if it is a simple question: shouldn't they be re-classified as well?

Thank you @fluzatayna for asking about this. Ideally I wouldn't just designate five sequences only, but I use a tree-based designation process, and one of the benefits of this is that there's additional support/confidence that all the designated sequences are really a clade. The tree sometimes lags behind the uploads to Gisaid, and there can be fewer members of a clade than expected. Nevertheless, pangolin should automatically recognise all descendants of that node as XBB.1.5.70 and assign them correctly as such, even though I haven't explicitly designated them. This has to be the case, otherwise we would have to keep designating new sequences as every pango lineage grew in size, which would be a very big job! If it turns out that pangolin is not assigning to XBB.1.5.70 properly, then I can designate some more to help clarify the assignments. The (UShER) tree can't be updated daily because it needs management (e.g. masking problematic sites, and manually resolving systematic errors with targeted designations) and time for the phylogenetics itself; this is the great work of @AngieHinrichs and others. There is also a check for sequence quality before designation, and this can also result in the apparent disappearance of some of the sequences found on Gisaid.