Assign clades using Troupin et al definitions

[kimandrews]

Description of proposed changes

Assigns clades based on widely used "major clade" definitions in Troupin et al. 2016.

Example tree is staged here

I would appreciate feedback on the following:

• Should we set "Clade" as the default coloring in auspice? For now I kept "Host" as the default coloring, because the clade definitions appear along the branches even when "Host" is used as default coloring, allowing the viewer to see both the host species and clade definitions at the same time.
• Should we modify clade names so they do not include geographic locations?

Future work could build on these changes to do the following for the rabies analysis:

• Assign subclades based on definitions in Troupin et al. 2016 and other sources
• Make a Nextclade dataset using major clade and/or subclade definitions
• Add a frequencies panel using clade definitions

Addresses #14

Related issue(s)

#14

Checklist

• Checks pass

[jameshadfield]

Example tree looks good - I agree with the decision to first focus on major clade. Have you run this build a few times to have confidence that the clades are consistently found?

Should we set "Clade" as the default coloring in auspice? For now I kept "Host" as the default coloring, because the clade definitions appear along the branches even when "Host" is used as default coloring, allowing the viewer to see both the host species and clade definitions at the same time.

👍 for the current defaults

Future work could... Assign subclades based on definitions in Troupin et al. 2016 and other sources

Worth playing around with but, as we discussed earlier, I'm not convinced that they're phylogenetically stable.

Future work could... Add a frequencies panel using clade definitions

Yes!

[kimandrews]

Example tree looks good - I agree with the decision to first focus on major clade. Have you run this build a few times to have confidence that the clades are consistently found?

Yes, I did run the build multiple times to check for consistency. Initially the bat clade had inconsistency in assignments across runs (i.e., sometimes some bat clade samples were not assigned to any clade). I noticed that the list of "unique mutations" along the branch leading to the bat clade changed a bit from run to run. So I looked to see which of those unique mutations were consistent across all runs, and only included those in the "bat clade" section of the clades.tsv. That resulted in consistent assignment of bat clade samples across runs.