Is B.1.355 a recombinant in sc2ts ARG?

[szhan]

Turakhia et al. (2022) suggest that B.1.355 is a recombinant of B.1.595 and B.1.371. We should check if this is the case in our sc2ts ARG.

Copying/pasting Figure 3 from the paper here.

[hyanwong]

They only have 3 mutations on the LHS, so wouldn't a recombination be unparsimonious according to our k=4 criterion.

I guess we could run the B.1.355 samples again using k=3, just to check?

[jeromekelleher]

Here's the match we get with k=3 for the first sample:

{"strain": "SRR13937758", "num_mismatches": 3.0, "direction": "forward", "match": {
"path": [{"left": 0, "right": 29904, "parent": 173}], "mutations": [{"site_id": 1485, "derived_state": "T", "inherited_state": "G", "site_position": 1495, "is_reversion": null, "is_immediate_reversion": null}, {"site_id": 7313, "derived_state": "C", "inherited_state": "T", "site_position": 7332, "is_reversion": null, "is_immediate_reversion": null}, {"site_id": 7811, "derived_state": "A", "inherited_state": "T", "site_position": 7831, "is_reversion": null, "is_immediate_reversion": null}, {"site_id": 27425, "derived_state": "T", "inherited_state": "C", "site_position": 27498, "is_reversion": null, "is_immediate_reversion": null}, {"site_id": 28281, "derived_state": "C", "inherited_state": "G", "site_position": 28362, "is_reversion": null, "is_immediate_reversion": null}, {"site_id": 28760, "derived_state": "T", "inherited_state": "G", "site_position": 28845, "is_reversion": null, "is_immediate_reversion": null}], "likelihood": 3.8146972656250014e-12, "cost": 6.0}}

which is the same as the match from the ARG:

   'hmm_match': {'mutations': [{'derived_state': 'T',
      'inherited_state': 'G',
      'site_position': 1495},
     {'derived_state': 'C', 'inherited_state': 'T', 'site_position': 7332},
     {'derived_state': 'A', 'inherited_state': 'T', 'site_position': 7831},
     {'derived_state': 'T', 'inherited_state': 'C', 'site_position': 27498},
     {'derived_state': 'C', 'inherited_state': 'G', 'site_position': 28362},
     {'derived_state': 'T', 'inherited_state': 'G', 'site_position': 28845}],
    'path': [{'left': 0, 'parent': 173, 'right': 29904}]},

There's 15 "B.1.355" samples in the ARG which I think neatly inherit from 19560 (I haven't checked they're all B.1.355, though).

So, I think the answer is "no" we don't think it's a recombinant.

[hyanwong]

Interesting that in the k=3 case there are still 3 mismatches @ 1485 (not in theirs), 1495, and 7313 (not in theirs). We don't have their 7831 mutation in that match.

Presumably we either don't have some of their recombinant parents, or we don't have their immediate recombinant child, otherwise we would have found their recombination event with k=3, right?

[jeromekelleher]

This is what we have as the root of our B.1.355 (I'm guessing):

[jeromekelleher]

Dunno, all seems pretty marginal.