KP.3 (w/the rare Q493E) has been my pick since I first noticed it emerging from numerous travel seqs from India. F456L & R346T are the typical stepwise immune-evasion mutations that, as @shay_fleishon noted, very likely impose a fitness cost. Q493E may be different. 1/

https://twitter.com/JosetteSchoenma/status/1785353182144610773

Q493E involves the rarest of all nucleotide mutations, C->G, and occurs at a key residue that we've seen very little action from of late. 493 mutations, however, are common in the Cryptics, usually Q493K I believe. (@SolidEvidence can correct me if I'm wrong on that). 2/8

We have a new record for mutations in a non-molnupiravir sequence. It's a BA.2.12.1 with >100 private mutations. There are 4 seqs from early April, all from the same patient. I'll discuss four interesting features it has in this 🧵. 1/23 #1) Reversions
Reversions are extremely rare. They almost never appear in circulating lineages. There are, however, a large number of reversions that are convergent in chronic-infection sequences. This one has more than usual. 2/23

What connects two regions on opposite ends of NSP12, a narrow slice of an obscure NSP3 region (DPUP/SUD-C), & a 3-AA sliver of nucleocapsid (N)? I have no idea, but I’m convinced there’s a link that could help reveal the inner workings of SARS-CoV-2. 1/120

I previously wrote a thread about the strange connection between ORF1a:4395-4398 and ORF1b:820-824 (NSP12_3-6 & NSP12_829-833). There is no known connection between these regions, & they are not close to each other in the NSP12 protein structure. 2/120

https://twitter.com/LongDesertTrain/status/1727852946291892367

Always nice to run across a possible function of a rare mutation that's shown up in multiple chronic-infection SARS-CoV-2 seqs. Thanks to an excellent paper by @TheMenacheryLab & @J_Paul_Taylor, I think I now know why N:L13P (a reversion) shows up. 1/6

https://twitter.com/TheMenacheryLab/status/1770959155093705059

They proved that the N:1-25 region, esp. the ITFG AA motif from N:15-18, is the essential element in N's ability to suppress the formation of stress granules (SGs) in cells, which capture & disable long viral RNAs & help organizing innate antiviral immune responses. 2/6

BA.2.86, a clear chronic infection-derived variant, has obtained near-total global dominance in the form of JN.1. This may have squeezed out any room for new CI-derived variants to take hold & spread, but they're still out there. A recent one had ~38 private spike mutations. 1/5

https://twitter.com/LongDesertTrain/status/1775807031649198306

BN.1.3 is a BA.2.75 descendant that emerged in Aug 2022, but much older variants still exist.

Alpha, for example, has disappeared from circulation, but it's not extinct—it's still evolving within an unknown # of hosts. On rare occasions, we catch a glimpse of this. 2/5

The spike mutation S50L is found in JN.1.
Before JN.1, it was extremely rare, though it was overrepresented in chronic-infection sequences. I think there's some evidence indicating that S50L might be an adaptation to the GI tract. 1/5 Virtually all SARS-CoV-2 sequences on GISAID come from conventional nasopharyngeal swabs. The small number labeled as coming from rectal/anal swabs, stool, or feces come almost entirely from China/Hong Kong, & are almost exclusively from February to April 2020. 2/5

An interesting BA.5.1 sequence turned up yesterday with ~32 spike mutations. As in all such cases, there's a lot of fascinating stuff happening, but I find three aspects of this one especially intriguing:
1) 4 spike reversions
2) ORF1a:K1795I
3) 2-nuc P9L, G446S 1/24 Spike Reversions
We're beginning to see old Omicron mutations cycling out, transforming from buzzsaw to anchor. Reversions K478T, A484E, & V486F are now among the most common mutations in BA.1/2/4/5 chronics. The D339G reversion here, however, is extraordinarily rare. 2/25

Interferon & T-cell escape is likely selected for in chronic infections. If high ACE2 affinity in SARS-CoV-2 evades IFN/T-cells, one might expect chronic variants to emerge w/high ACE2 affinity. And indeed, we've seen exactly that w/BA.2.86 & BA.2.87.1 1/5
h/t @Nucleocapsoid

https://twitter.com/LabWaggoner/status/1769809271615570283

"High receptor-binding affinity virus particles exhibited increased resistance towards IFN-I-mediated inhibition of viral entry & replication... eventually leading to impaired T cell immunity and prolonged virus load."
2/5

I love it. We now understand that even if a unique, divergent lineage doesn't spread far, it represents a golden opportunity to learn things about this virus we wouldn't otherwise be able to. And such lineages have often foreshadowed the mutations of future variants. 1/10

https://twitter.com/dfocosi/status/1767841861425643957

In some ways, BA.2.87.1 is reminiscent of BJ.1, an obscure BA.2 chronic-infection lineage that emerged in India but hardly grew at all. It vanished quickly, but >75% of its genome lived on in XBB (a BJ.1/BA.2.75 recombinant) & went on to dominate globally. 2/10 image: @PeacockFlu

Great new preprint on BA.2.87.1. Neutralizing antibody results & reduced lung cell infectivity compared to JN.1 is encouraging. Higher cell-cell fusion less so but not surprising given the ability of NTD deletions to facilitate fusion. There's also one puzzling finding. 1/23

https://twitter.com/MIID_alex/status/1763392811423141914

In case you forget what BA.2.87.1 is, here's a previous thread on it. BA.2.87.1 has not spread at all since then, so it is not a major concern at the moment.

2 mistakes I made: R346T is not in BA.2.87.1, & I should've included ∆69-70. 2/23

https://twitter.com/LongDesertTrain/status/1753195705638760582

This Cryptic in Lausanne, Switzerland, disappeared for 2 years, & now returns in nearly unrecognizable form, w/a spike RBD resembling no known virus. It's clearly changed receptors. It seems we've barely scratched the surface of the potential changes SARS-CoV-2 can undergo. 1/4

https://twitter.com/SolidEvidence/status/1762255568117387576

My extremely amateur attempt to align these Cryptics with some SARS-CoV-2-like Clade 1a sarbecoviruses and SARS-CoV-1-like Clade 1b sarbecos. 2/4

Fortunately, use of the ineffective, mutagenic Covid drug molnupiravir (MOV) has declined. Unfortunately, several countries are still using it, including the US, Japan, South Korea, & Israel. A cluster of three JN.1 sequences with ~60 new mutations just turned up in Israel. 1/8 The number of mutations alone marks these as MOV creations, but the mutation types also confirm that diagnosis. All 60 mutations are transitions, with 32% of those G->A, the most distinctive MOV mutation.
The three sequences differ slightly, possibly due to coverage issues. 2/8

This could have the record for private spike mutations, surpassing even BA.1 & BA.2.86. ~35 spike AA mutations—& likely many others hiding behind dropout at S:447-471. Singlet in a country w/good surveillance, so unlikely to transmit. I'm super busy, so a few short remarks. 1/15 First, of the nucleotide mutations in amino acid (AA) coding regions, 61/71, or 85.9%, are non-synonymous, i.e. they cause a change in the AA. This is extraordinarily high and indicates strong positive selection—i.e. selection for advantageous mutations. 2/15

Time for more Tales from the Cryptics.

This one's a doozy. Has to be the wildest RBD yet.

It's hard to convey just how crazy this one is, but in this 🧵 I'll share my initial thoughts one some of these RBD mutations. (RBD = receptor-binding domain of the spike protein)
1/22

https://twitter.com/SolidEvidence/status/1756467929242022203

I've analyzed 1000's of chronic-infection sequences, yet this Cryptic has 9 mutations I've never seen in any of those: G413R, T415E, Y421F, N440T, V445N, G446K, Y449T, K462Q, and G482T—*in the RBD alone.* Two others I've only seen once. 2/22

Longer 🧵 on BA.2.87. So far, it has not grown quickly or spread far geographically, so its future is murky. It could be a flash-in-the-pan that soon disappears or it could mutate to become more fit & challenge JN.1. My summary of its private mutations is below. 1/22 This is *not* to be confused w/the equally amazing BA.2 singlet discovered by @BorisUitham. While these 2 astoundingly divergent BA.2s are completely different, they do share one remarkable similarity: total loss of the same spike NTD disulfide bond. 2/22

https://twitter.com/LongDesertTrain/status/1752320762537525489

Phenomenal 🧵 by @Tuliodna describing a wild new BA.2 lineage circulating at low levels in South Africa.
Incredible find by one of the best & most valuable teams in the world. We're all in their debt. Thank you, @Dikeled61970012, @nicd_sa, @DarrenM98230782, @houzhou, & team.

https://twitter.com/Tuliodna/status/1752996491948327126

Will this new BA.2 sputter and disappear without making much impact? We've seen many chronic-infection BA.2 saltation variants do exactly that: BA.2.83, DD.1, BP.1, BA.2.10.4, + multiple undesignated ones (exceptionally divergent BA.2s in Chile & Ukraine come to mind). 2/

We see a lot of BA.2* these days: ~100% of cases have been due to BA.2-derived variants in the past year. But what ever happened to BA.1?
It’s still around, just not transmitting… much. The BA.1 below, w/32 additional spike mutations, was transmitted at least once. 1/15 Another recent variant also had 32 spike mutations relative to its ancestor: BA.2.86 (which later became JN.1). This is not the first extremely divergent, chronic infection-derived BA.1* sequence we’ve seen. But it may be the first we’ve seen transmitted. 2/15

I'm not 100% sure due to uncertainties in the sequencing, but I think something remarkable has happened in JC.5.1 (an XBB.1.41 branch that has S:Q173K, L335S, R403K, K478R, S486P, & N:H300Y).
There's a large ORF7a deletion that leads to ORF7a-7b fusion—and more. 1/8 This deletion causes a change in the reading frame, i.e. a frameshift. This not only leads to fusion of ORF7a-7b but also completely changes the amino acid content of the last 20 AA of ORF7a. 2/8

So the deer-derived sequence I described in the thread below has a companion—from a different US state. They share 35 mutations, so they are clearly related. But the shorter branches leading directly to each seq have 19 and 24 mutations, so they differ by 43 nuc mutations. 1/13

https://twitter.com/LongDesertTrain/status/1743099275062517837

These are BF.11's, a lineage that hasn't circulated in humans for ~1 year. There are 2 possibilities:
1) This deer variant is circulating at low levels in humans in the US Midwest
Seems extremely unlikely, esp w/great recent surveillance in Minnesota. @CIDRAP @mtosterholm 2/13

Delta's still hanging around—and packing heat, in the form of 30+ spike mutations. As w/other recent Delta singlets, this is very likely 1 chronically infected person‚ not a circulating variant. But it's a healthy reminder that the prospect of a Delta comeback always looms. 1/15 A few aspects of this Delta worth noting:
• About 9% of the genome lacks coverage, including >10% of spike, so the mutations indicated are an undercount. Hopefully, we never see this particular virus again and will never know what's hiding behind the dropout. 2/15

Quick 🧵on a JN.1 sublineage w/an unusual mutation, now designated JN.1.5 by the indefatigable @CorneliusRoemer. It seems to be the only large JN.1 sublineage that enjoys a slight growth advantage over the OG JN.1, though @siamosolocani may correct me on that. 1/27 What's unusual is that this has a 2-nucleotide mutation in ORF1b:V1273T (NSP13_V348T). Two-nuc mutations are rare, yet this one has also appeared in three XBB* lineages in the past 6 months (FL.15.3, XBB.1.16, & JG.3). 2/27