nextstrain / auspice

Web app for visualizing pathogen evolution
https://docs.nextstrain.org/projects/auspice/
GNU Affero General Public License v3.0
290 stars 159 forks source link

Show mutation split into homoplasic/unique/reversions also on terminal branches, not just internal ones #1752

Closed corneliusroemer closed 4 months ago

corneliusroemer commented 4 months ago

Context

The Auspice feature to split mutations on branches into: homoplasic, unique, and reversions is very useful.

I noticed that this feature is not available for terminal branches. I can't see a reason why we shouldn't use the same feature for those branches, it's just as useful there.

See internal:

image

Vs. terminal:

image
FedeGueli commented 4 months ago

@corneliusroemer @ryhisner i think it was available until friday and now it changed and it has been removed.. if this is correct it is really a pain for us regularly checking singlets everyday maybe @ryHisner could add his point

ryhisner commented 4 months ago

Terminal-branch mutations have always been visible. They just suddenly disappeared a day or two ago, and I REALLY miss them. It's hard to express just how much harder it is for me to do what I usually do without being able to view them. I've already wasted hours trying to sort out the mutations on long branches—something I used to be able to do quickly because shift-clicking the branch showed them immediately. And you could see a preview of the mutations just by hovering over the branch. If it was a relatively short terminal branch (<8 nuc mutations), you could see all the mutations this way.

ryhisner commented 4 months ago

I noticed that even the 21L tree has changed—you can no longer view the mutations on terminal branches by shift-clicking. Since there's no option to label branches with nucleotide mutations, it's extremely difficult to tell what synonymous (or non-synonymous ORF9b) mutations a terminal branch has.

image
corneliusroemer commented 4 months ago

Thanks guys for suggesting this is a regression and that this happened recently.

It should be easy to revert for @jameshadfield - as it's not in the changelog I don't think it was intended.

FedeGueli commented 4 months ago

Thanks guys for suggesting this is a regression and that this happened recently.

It should be easy to revert for @jameshadfield - as it's not in the changelog I don't think it was intended.

Thank you @corneliusroemer @jameshadfield

corneliusroemer commented 4 months ago

I can confirm this is indeed a regression introduced between 2.51.0 and 2.52.0

This is 2.51.0 (npx auspice@2.51.0):

image

This is 2.52.0 (npx auspice@2.52.0):

image
FedeGueli commented 4 months ago

I can confirm this is indeed a regression introduced between 2.51.0 and 2.52.0

This is 2.51.0 (npx auspice@2.51.0): image

This is 2.52.0 (npx auspice@2.52.0): image

Thank you Cornelius as variant spotter that slows very much or disable our work efficiency . I know nextstrain is much more important than for our goals, but it is also a time where there aren't many people doing spotting systematically so it would be very important this will be fixed soon. thank you very much for your help.

ryhisner commented 4 months ago

Is a fix for this in the works? I'm trying workarounds to be able to figure out all the mutations on terminal branches, but they all require a lot more time.

corneliusroemer commented 4 months ago

Using auspice/auspice.us 2.51.0 would work - though you'd have to host them locally AFAIK

jameshadfield commented 4 months ago

I've fixed this in PR #1753. I'll merge shortly and make a new Auspice release (2.52.1) and push this over to nextstrain.org. It should be available in the next few hours if all goes smoothly.

FedeGueli commented 4 months ago

I've fixed this in PR #1753. I'll merge shortly and make a new Auspice release (2.52.1) and push this over to nextstrain.org. It should be available in the next few hours if all goes smoothly.

Wow! great news thank you very much!

jameshadfield commented 4 months ago

This issue has been closed because the PR has been merged, but the fix is yet to propagate to nextstrain.org. I'll write a message here when this happens.

For those interested in what happened, a recent PR (#1749) inadvertently removed the distinction between terminal branches and tips, which resulted in hovering or shift-clicking on terminal branches showing the wrong information. The intended behaviour is for hovering/shift-clicking on a branch to summarise the changes on that individual branch, whereas hovering/clicking on the tip should show information about the specific strain and (e.g.) collect all mutations between the tip and the tree root.

jameshadfield commented 4 months ago

Changes released with Auspice 2.52.1, and nextstrain.org is updated. If you are still getting the old version (as shown in the page footer) please try refreshing the page holding the shift key.

FedeGueli commented 4 months ago

This issue has been closed because the PR has been merged, but the fix is yet to propagate to nextstrain.org. I'll write a message here when this happens.

For those interested in what happened, a recent PR (#1749) inadvertently removed the distinction between terminal branches and tips, which resulted in hovering or shift-clicking on terminal branches showing the wrong information. The intended behaviour is for hovering/shift-clicking on a branch to summarise the changes on that individual branch, whereas hovering/clicking on the tip should show information about the specific strain and (e.g.) collect all mutations between the tip and the tree root.

Thank you for the explanation and for the tip !

ryhisner commented 4 months ago

I've fixed this in PR #1753. I'll merge shortly and make a new Auspice release (2.52.1) and push this over to nextstrain.org. It should be available in the next few hours if all goes smoothly.

You guys are awesome! Thanks so much!

corneliusroemer commented 4 months ago

Thanks @jameshadfield!