COVID-19 cluster with unusual clinical presentation in France

[Simon-LoriereLab]

Description request by the French National Reference Center for viruses of respiratory infections headed by Pr. S. van der Werf

Sub-lineage of: B.1 Earliest sequence: 2021-01-29 Most recent sequence: 2021-02-15 (more sequencing ongoing) Countries circulating: France Genomes Table_Lannion.tsv.txt

Evidence A cluster of COVID-19 cases with inconsistent detection by RT-PCR in nasopharyngeal swabs at the time of onset of symptoms compatible with COVID-19 (including severe/lethal cases) is under investigation in Bretagne France. More sequences are under way, but most samples are very difficult to sequence (low viral load with the usual sampling methods) The genomes are characterized by a large number of mutations:

Spike_D215G,Spike_H655Y Spike D215G D614G G142del G669S H66D H655Y N1187D Q949R V483A Y144V E F20L T30I M H125Y N T325I NS3 Q57H NSP2 T85I NSP3 N506S S1443Y T820I NSP4 Y397H NSP5 A260V NSP6 L37F NSP12 P323L Q822R NSP14 L157F NSP16 K277R T140I

tree_Lannion_20C_B.1.pdf tree_Lannion_20C_B.1.nex.txt

A Pango designation would greatly help exchanges on this cluster!

[AngieHinrichs]

GISAID sequence France/IDF-HMN-21022230289/2021|EPI_ISL_1110211|2021-01-29 also clusters with these:

[AngieHinrichs]

FWIW the three BRE-IPP* sequences and IDF-HMN-21022230289 also have a 29bp frameshift deletion in ORF6, 27267-27295 (or 27268-27296 depending on which way your aligner rolls).

Also, the deletion of 3 bases in spike appears to me to be not G142del, but rather Y144del or Y145del (depending on which way the aligner rolls): 21991-21993 ... 21994-21996. B.1.1.7 has the same deletion.

[rambaut]

This doesn't currently qualify as a new lineage under the Pango designation criteria. It needs to a have at least 5 cases and at least 1 internal node. If more cases arise then it will be revisited. I am marking this as monitor but leaving it open.

[AngieHinrichs]

Still no internal node, but HT Anna Niewiadomska (JCVI) for pointing out two more sequences that cluster with these, EPI_ISL_1259297 and EPI_ISL_1259307. EPI_ISL_1239370 does too. That brings it up to 7:

France/IDF-HMN-21022230289/2021 EPI_ISL_1110211 2021-01-29
France/BRE-IPP03921/2021    EPI_ISL_1111064 2021-02-15
France/BRE-IPP03808/2021    EPI_ISL_1118892 2021-02-10
France/BRE-IPP03809/2021    EPI_ISL_1118893 2021-02-10
France/BRE-IPP04583/2021    EPI_ISL_1239370 2021-03-02
France/BRE-IPP04233/2021    EPI_ISL_1259297 2021
France/BRE-IPP04392/2021    EPI_ISL_1259307 2021-02-26

Interactive view: https://nextstrain.org/fetch/hgwdev.gi.ucsc.edu/~angie/pango-designation-31.json?branchLabel=nuc%20mutations&c=pangolin_lineage&label=nuc%20mutations:A6653G

It looks like my tree has France/BRE-IPP03809/2021|EPI_ISL_1118893 before the node with A22206G, but then UShER is placing the same sequence when uploaded after the node with A22206G -- ah OK, BRE-IPP03809 has 'N' at 22206 so UShER made different decisions in different contexts about whether to infer ref or alt for the missing base. Test case for Yatish's matOptimize.

However, the relationship of France/BRE-IPP04615/2021|EPI_ISL_1262772 to the sequences in that cluster is... interesting. It has 24 of these 26 mutations in common with the cluster (it appears not to have G25563T/ORF3a:Q57H and C1059T):

C241T > C14408T > A23403G > C3037T > G25563T > C1059T > C5884T > C1973T, C5178T, C7047A, T9743C, A15905G, C18508T, C21077T, A21488G, C21691T, A25121G, T26304G, C26333T, C26895T, C27128T, T27757A, T28297C, C29095T, C29247T, C29580T

-- and A22206G (S:D215G), which due to BRE-IPP03809's N and the timing of adding sequences to UShER, appears to happen later for the cluster but probably happened before the split with BRE-IPP04615.

It also shares the deletion of 21991-21993 (S:Y144del or S:Y145del) with the cluster.

But then the cluster has these 9 mutations that BRE-IPP04615 does not have:

A4236G, C10252T, C21758G (S:H66D), G21987T (S:G142V), T23010C (S:V483A), C23525T (S:H655Y), G23567A (S:G669S), A24408G (S:Q949R), del27267-27295 (ORF6 frameshift/truncation)

And BRE-IPP04615 has these 15 mutations that do not appear in any sequence in the cluster, but have 1 in common with B.1.1.7 (bolded):

C5388A ORF1a:A890D C5986T C8090T T12496C T16176C A17615G del 21765-21770 S:69-70del A23063T S:N501Y C23604A S:P681H C23709T S:T716I G24914C S:D1118H C27972T ORF8:Q27stop G28048T A28111G G29717A

B.1.1.7 mutations not in BRE-IPP04615: C913T C3267T T6954C del 11288-11296 (ORF1a 3675-3677) C14676T C15279T C23271A (S:A570D) T24506G (S:S982A) del 28271 28881-28883 GGG -> AAC (N:203-204) C28977T (N:S235F)

I don't see a tidy breakdown into a small number of regions of the genome that are clearly like the cluster vs. regions that are clearly like B.1.1.7. I hope someone who knows much more about sequencing and viruses than I do can look at France/BRE-IPP04615/2021|EPI_ISL_1262772.

[rambaut]

Our general policy is to treat any signal that is in one genome sequence only with caution as will often be due to sequencing issue (i.e., contamination) or mixtures present in the sample. The fact that there is no simple split between B.1.1.7-like and cluster-like mutations suggest a mixture of amplicons from the two source variants.

[rambaut]

I am definitely going to call IPP04615 as a mosaic (the two mutations in ORF8 downstream of the Q27 stop codon are unlikely to have occurred independently - even though Q27 stop has occurred multiple times). Could be a recombinant of course but the switching back and forth makes it more likely a mixture with a consensus that is mosaic - the raw data needs inspection.

[Simon-LoriereLab]

Thanks to both for looking into this (it has hijacked a bit the unusual cluster thread, we are indeed generating more sequences and will post an update when/if the threshold is reached). The raw data appears partially mixed but we are checking this in more details, and will surely re-sequence, looks like a potential co-infection.

[rambaut]

Hi Etienne - thanks for looking into this. One other thing about the cluster - is it associated with a specific outbreak (I heard it was a hospital outbreak) or is it spreading more generally? The long unsampled branch connecting to genomes from the first half of 2020 is suggestive of an introduction from an area with limited genomic surveillance - is there any evidence of that?

[Simon-LoriereLab]

Hi Andrew, for now it is indeed a specific outbreak in a hospital(+care facility for the elderly) with cases in patients and HCW. The sequence outside of the region has a direct epi link. The difficulty to confirm these cases by RT-qPCR on nasal swabs possibly mean that it will be a challenge to capture it outside of the severe/hospitalized cases, but a lot of epi work is ongoing. We currently have no evidence of introduction from another area. More analysis is on the way, but it might not be possible to get a proper answer on this.

[Simon-LoriereLab]

Dear colleagues, here is an update on the cluster, with more sequences on the way.

2021-03-23_update_cluster_low_detec.nex.txt

[erikalmecdc]

Dear all,

This is a very high profile variant that could explode into a new VOC if it turns out to spread further. It would be very nice to get a lineage designation to be used instead of the now circulating designation "Brittany variant".

[chrisruis]

Hi All, we've now assigned this lineage as B.1.616. It should appear in the next update. Thanks!

[Simon-LoriereLab]

Dear Chris, dear colleagues, thank you for your help with this. It will be very useful.

[AngieHinrichs]

Looks like there are 4 additional sequences in this cluster, from as recently as 15 March:

hCoV-19/France/BRE-HCL021036256901/2021 EPI_ISL_1318289 2021-02-10
hCoV-19/France/BRE_8600536377/2021  EPI_ISL_1292807 2021-02
hCoV-19/France/BRE-IPP05435/2021    EPI_ISL_1381829 2021-03-10
hCoV-19/France/BRE-IPP05485/2021    EPI_ISL_1381830 2021-03-15