ENH: Display ambiguous mutations in gene/nuc view tooltip

[corneliusroemer]

Ambiguous nucleotides are currently displayed only in the tooltip of the corresponding column:

In gene view, ambiguous bases turn the corresponding AA into an X. It would be nice, if the X tooltip showed corresponding ambiguous nucleotides, like AA mutation tooltips show the mutated nucs.

Current:

Something like this would be cool, potentially even with a range of possible amino acids that is compatible with the ambiguity.

[ivan-aksamentov]

The translation of ambiguous nucs is a cool feature we discussed with Richard even in the early days of Nextclade. It can be implemented quite easily, by modifying the decoding map. Someone needs to actually build the map and put the additional triplet-to-aa entries above the catch-all case which resolves to AA X. It would be to error-prone to do manually, but perhaps one could take some known good Python lib with translation capabilities, and print the required Rust entries using a small script.

The coolness of this I see in that it will reduce the number of X, and salvage some amount of useful info even from mediocre data. Which is always nice and honorable, because sequencing data is precious.

Regarding showing how ambiguous nucs and AAs are linked - it's trickier. We currently don't track why a codon becomes an X and there are multiple reasons why a codon can do that. Also X are tracked as ranges, and within a range there could be multiple reasons by itself, so some deeper code changes will be needed.

[corneliusroemer]

Just wanted to open a duplicate then found this: I was surprised that we don't indicate ambiguous nucleotides in the nuc view. Quite unexpected, I think we don't need to overcomplicate things: just add them like a single missing nuc.

[ivan-aksamentov]

@corneliusroemer showing nuc markers is easy. I can add that. What colors would you like to see? We already have 3 shades of gray for deletions, Ns and unsequenced. And if not gray, then I guess some random pale colors?

However calculating correspondence between aa and nucs is generally a more complex task. Plus, not all X come from ambiguous nucs.

We've improved our decoding recently to handle some of the ambiguities, so, hopefully, there should be less Xs coming from non-ACTGNs.

[corneliusroemer]

Some shade of grey would be the obvious choice I think. Doesn't matter so much what it is, as long as there's something that can be hovered to see extra information.

I don't think we need to add anything related to AA there, just showing that there might be something in nuc view is good enough. Just putting the relevant entry from the table tooltip into the marker tooltip would be great:

While we're at it, it would be neat to maybe add what the various ambiguity codes stand for. This is a simple lookup:

R = A or G
K = G or T
S = G or C
Y = C or T
M = A or C
W = A or T
B = not A (C, G or T)
H = not G (A, C or T)
D = not C (A, G or T)
V = not T (A, C or G)

So the tooltip could read:

Ambiguous 
11514: Y (C or T)

[corneliusroemer]

Would be great to add this @ivan-aksamentov - this would have really helped me some time investigating this case: yatisht/usher#354 - not showing ambiguous nucs/aa would be a big win for the sequence viewer. Color doesn't really matter, we can always tweak later.

[ivan-aksamentov]

@corneliusroemer prototype is here: #1273