XBB-like ARTIC v4.1 Dropout

[ktmeaton]

There appears to be sequencing dropout (Ns) in the region of 22248:22428 in XBB. This leads to the breakpoints to be off by 141 nucleotides (expected 22332, observed 22191). This regions is exactly the start/end coordinates of ARTIC v4.1 primers 73,75 and is only observed in samples prepped with ARTIC (as opposed to Midnight/Freed). The simplest explanation would be amplicon dropout.

Some options I'm considering:

• [ ] ~Relax the breakpoint error tolerance from 50 nucleotides to 200 nucleotides.~
• [x] Add separate entries in resources/breakpoints.tsv for XBB_ARTICv4.1, and then remap this to XBB in scripts/linelist.py.

I prefer the second one, because I don't have to re-optimize/test parameters for all previous recombinants which have very similar breakpoints.

[ktmeaton]

There typically aren't any substitutions in this area, the last one before is T22200G and the first one after is G22578A

[ktmeaton]

This style of dropout (Ns 20677-22537) is being labeled as a false positive, when I would prefer it to be XBB-like. Why is this a false positive?

Note: This is sequenced with Midnight/Freed, not ARTIC

[ktmeaton]

In the latest version (v0.6.2 in development), this same sample is classified correctly as XBB.1.5-like. I'll attribute this to improvements in the latest version of nextclade aligner or dataset.

OR! C20235T has now snuck into being a diagnostic mutation for XBB (`0.370) with the latest allele frequency update.