Post #1861707

2023-09-20 11:10 UTC

V2 of my preprint with PhD student Kim Verity is now online: https://www.biorxiv.org/content/10.1101/2023.07.17.549219v2 tldr: If you want to assess the shape (e.g. balance, diversity) of a rooted tree then generally you&#39;re better off using our new indices than Sackin&#39;s index, phylogenetic entropy, etc. The article is long mainly because we provide lots of examples. I plan soon to write a shorter #mathonco blog post explaining why these indices are better than prior approaches. Meanwhile, here&#39;s a poster.

Replies (2)

@robjohnnoble@ecoevo.social 2023-09-20 11:12

Unlike currently popular phylogenetic diversity, phylogenetic entropy, and tree balance indices, our definitions assign interpretable values to all rooted trees and enable meaningful comparison of any pair of trees. For example, this preprint figure compares the shapes of trees representing the evolution of HIV and of Uralic languages. If we ignore branch lengths then the trees are indistinguishable (a-d). But our new indices accounting for branch lengths capture important differences (e-h).

Open ##2713949
@ryneches@ecoevo.social 2023-09-22 13:32

@robjohnnoble@ecoevo.social This is so cool! Have you considered applying this approach to index design to co-phylogenies? I've been working on supervised learning for co-phylogeny classification problems for a long time, and the biggest hurdle has been establishing some kind of ground-thruth for the classifier output. Metrics on trees are often frustratingly unrelated to the simulation parameters used to make the trees in the first place, which bodes poorly for their application to trees inferred from data.

Open ##2713954