XN not a recombinant in the ARG

[hyanwong]

It appears as if we conclude that a few Pango-designated recombinants are not the result of recombination in the ARG. This includes XB (which we have always have trouble with) but a better example is XN (below; XP is another, similar example). I wonder if this is because we are excluding some mutations (indels?) that are being used for the recombinant designation?

[jeromekelleher]

XP has a characteristic deletion, which we've always missed: cov-lineages/pango-designation#481

Here's the pango designation issue for XN: cov-lineages/pango-designation#480

[jeromekelleher]

The origins of XN look pretty clean to me, I'm not sure it really is a recombinant. The first sample has this copying path:

'hmm_match': {'mutations': [
     {'derived_state': 'C',      'inherited_state': 'T',      'site_position': 2790},
     {'derived_state': 'G', 'inherited_state': 'A', 'site_position': 2832},
     {'derived_state': 'A', 'inherited_state': 'G', 'site_position': 10986}],
    'path': [{'left': 0, 'parent': 983443, 'right': 29904}]},

So only three mutations from 983443.

That node looks like a pretty solid inference to me with 107 children and 186 exact matches:

[hyanwong]

It would be good to point out that XN really isn't a recombinant, on closer inspection.