ENH: Make it clearer when Pango calls are expected to be wrong

[corneliusroemer]

It doesn't seem entirely clear to users when Pango assignments are trustworthy and when not, see #760

One could for example add parentheses about the Pango assignment when the quality is bad to make the reduced confidence clear.

Or maybe the tooltip needs to be updated to give explicit hints on when to trust it and importantly when not.

[AnonymousUserUse]

Is this issue already tackled, and if yes, is it visible in CoV-Spectrum?

I wonder whether the assignment of XAP from pangoLEARN (80 sequences with coverage >0.9) or Nextclade (1811 sequences with coverage >0.9) is more correct and hope that I can see the answer for those kinds of questions in certain website.