Open huzuner opened 2 months ago
wm75
commented
2 months ago If it helps, even nextclade (at https://clades.nextstrain.org) itself doesn't classify your sequence as 21E but as 20B.
I don't know why that would be the case since, indeed, you seem to have put the 21E clade's variants into the sequence, but this then becomes a question to ask at https://github.com/nextstrain/ncov.
huzuner
commented
2 months ago Yes, that's right. It rather seems to be a problem related to Nextstrain. Thanks for your insight!
AngieHinrichs
commented
2 months ago Hi @huzuner, thanks for sharing your example 21E. I think the issue, probably for both pangolin and nextclade, is that your 21E sequence has only a subset of the many mutations found in most 21E sequences. It looks like your sequence has these mutations (relative to the NC_045512.2 reference):
C12049T, C14408T, T23341C, A23403G, C23604A, T24187A, G24836G, G28881A, G2882AHowever, in the UShER tree, these mutations are found on the path to the branch with 21E, i.e. they are present in most 21E sequences:
C241T, C3037T, A4926G, C7564A, C8139T, G9203A, T9867C, C11308A, C12049T, C12053T, C14408T, C17074T, C17339T, A22356G, G23012A, A23063T, T23341C, A23403G, C23604A, T24187A, G24836A, C24863T, G25088T, A27897C, G28881A, G28882A, G28883C, C29724TSo your sequence is mostly reference, with a small subset of the mutations actually found in 21E. https://github.com/nextstrain/ncov/blob/master/defaults/clades.tsv generally has only enough mutations to distinguish the Nextstrain clades from each other, and Nextstrain's augur pipeline identifies clades on a tree using an algorithm that expects to use only a subset of the mutations as input. But an actual 21E sequence really should have many more mutations than just the ones in clades.tsv. I believe nextclade does not use clades.tsv when assigning clades to sequences; it uses a tree, similar to pangolin/usher.
huzuner
commented
2 months ago Hi @AngieHinrichs, thank you so much for your detailed insight into my problem. I really appreciate it.
That now explains a lot why my experiments failed all the time with Pangolin and Nextclade predictions carried out on simulation experiments. I mentioned this here earlier today: https://github.com/nextstrain/ncov/issues/1107#issuecomment-2217961027 The problem for me is that I'm having difficulties finding a consistent resource for Nextclade sequences. I would be glad if you state your opinions on that as well.
Thank you in advance!
AngieHinrichs
commented
2 months ago @corneliusroemer maintains a set of Pango lineage consensus sequences that should work much better with nextclade and pangolin: https://github.com/corneliusroemer/pango-sequences
huzuner
commented
2 months ago Is there a resource like this for Nextclade/Nextstrain clades?
Hello,
I have been using Pangolin to evaluate my tool on a simulated dataset.
However, I cannot get the lineage that I expect with pangolin. I would be glad if you could have a look why that could be the case.
I have reference sequences for each Nextclade clade that I generated myself by changing the nucleotide positions (https://github.com/nextstrain/ncov/blob/master/defaults/clades.tsv) when compared to the reference sequence.
Below is an example for 21E:
I then basically use e.g. 21E as input to pangolin with default parameters:
pangolin 21E.fasta --outfile 21E.tsvHere is the resulting output:
taxon,lineage,conflict,ambiguity_score,scorpio_call,scorpio_support,scorpio_conflict,scorpio_notes,version,pangolin_version,scorpio_version,constellation_version,is_designated,qc_status,qc_notes,note 21E ,B.1.1.28,0.0,,,,,,PUSHER-v1.27,4.3,0.3.19,v0.1.12,False,pass,Ambiguous_content:0.02,Usher placements: B.1.1.28(1/1)B.1.1.28 is equal to 20B using the resource for mapping Pango lineages to Nextclade nomenclature: https://hgdownload.soe.ucsc.edu/goldenPath/wuhCor1/UShER_SARS-CoV-2/
Do you have any clue what could be the reason that Pangolin does not output a Pango lineage that corresponds to 21E?
Thank you in advance for your insights!
Best, huzuner