`Rogue::ColByStability` incorrectly marks the positions of completely stable taxa as unstable

[HS6986]

Dear Professor Smith,

This issue has been mentioned at ms609/TreeTools#168, but I will describe it here as a reminder.

Rogue::ColByStability seems to sometimes incorrectly mark the positions of completely stable taxa as unstable, as exemplified by the case of Calvadosia and Haliclystus in ms609/TreeTools#168. A similar condition can be reproduced using the minimally reproducible example I provided there.

Thank you for your time and support.

[ms609]

Thanks, I'll investigate – I wonder whether this is related to the former issue.

[ms609]

I think this is displaying the anticipated behaviour. The trees from the file are read as rooted trees, and the graph geodesic counts two edges at the position of the root, where only one would be counted if the trees were not rooted. With

trees <- read.nexus("treeFile-02.txt")
Rogue::TipInstability(trees) # gives high values for taxa on 'other side' of the root
Rogue::TipInstability(ape::unroot(trees)) # reports zero for all taxa

Whilst investigating, I found a rounding error that caused ColByStability to allocate random colours when all taxa were perfectly stable due to floating point imprecision.

[HS6986]

Thank you for your reply and for fixing the bug you found.

It looks like this was not actually an issue. I apologize for the inconvenience.

Does using Rogue::X() require trees that are either unrooted or appropriately rooted? I have overlooked this point. If true, it may be a good idea to mention it explicitly in the documentation of Rogue.

I appreciate your time and support.

[ms609]

It's a good question – for TipInstability the answer is that trees that are supplied as rooted are treated as such, whereas for e.g. QuickRogue the position of the root is irrelevant to the calculation. I agree that this would be useful to document, so I'll reopen this issue as a reminder to do so – or you are welcome to propose an edit yourself.