Nextclade GII.4 assignment

[j23414]

Description:

Migrating as a follow up to an initial Nextclade dataset, diagnose and fine-tune GII.4 and variant typing. See highlighted nodes here.

There are a few different potential solutions to pursue on different follow up dev branches:

• Change the reference sequence from GII.4 to ???
• Check if the inferred ancestor provides a reasonable VP1 coding sequence (extremely unlikely due to diversity of norovirus and recombination)
• Adjust the alignment presets
• Stricter unique mutation cutoffs
• Directly specify the types of these samples, and add them to the Nextclade dataset scaffold tree
• Potentially develop a GII.4 Nextclade dataset for variant/cluster/subvariant classification

[joverlee521]

Change the reference sequence from GII.4 to ???

I remember we chatted about this and there were many reference seqs on NCBI...Did you test out whether this helps with the issue?

[j23414]

Tried switching to the reference sequence on NCBI for GII.4 on this branch main...update-nextclade-reference

Then went in the direction of trying to root the tree on an outgroup on this branch so that we could have an "unclassified root"
main...align-nextclade

Norovirus is diverse enough that the internal inferred root sequence had indels (which breaks Nextclade). I'm open to suggested paths forward though

[j23414]

We need [Edit: I was aiming at] a high quality, well rooted VP1 and RdRp gene trees (not the full 14). So this is really a search for a VP1 and RdRp tree with:

• unclassified root
• clearly separated group, type, variant
• tuned nextclade alignment parameters to handle diversity of norovirus genome (many indels) so adjust against the main reference
• perhaps allow for multiple norovirus references to call mutations against (stretch goal)

[j23414]

From comment

That's fair. I'm fine with adding explicit warnings in about the false positives in the README and description.