File traitsdivision.mugration_model

General
1

Hello, checking the files in the results folder I found the file traitsdivision.mugration_model.txt
I have searched how to understand these tables but I can’t find anything in the forum, can someone help me to interpret the results of said file?

2

Hi @juan_dc. These files are output as part of running “augur traits”, which is effectively a wrapper around TreeTime’s “mugration” model. In a trait model, the rate of transition from states A→B, A→C, etc… is estimated. The mugration_model.txt output is pairwise matrix listing these rates. These pairwise rates should be largely comparable to pairwise rates estimated by BEAST using a standard phylogeographic model.

In addition, you get a JSON file that gives estimated locations of internal nodes in the tree and relative probability of these locations.

1 Like
3

Thank you very much for the reply.
Kind regards from Colombia

Juan D. Ceron

4

I am seeking additional clarification regarding the interpretation of transition rates. Given that these rates are similar to those used in BEAST, can they be understood as the number of transitions per unit of time, per state? For example, if my states represent avian hosts (e.g., goose, chicken) and I am analyzing a viral gene of influenza with full sampling dates available, would the unit be transitions per year? In my asymmetric transition matrix, if the goose-to-chicken rate is 1.5 and the chicken-to-goose rate is 0.5, can I interpret this as 1.5 transmissions per year from geese to chickens and 1 transmission every two years from chickens to geese? I appreciate any clarification on this. Thank you!

1 Like
5

Hi @mazeller
the interpretation of the rates will depend on the tree that is used to infer the mugration model. The rates themselves will be the rate at which state changes occur along the branches in the tree using the same units in which the branch lengths are given. If the units of branch lengths are for example subsitutions per site, then inferred rates will be relative to the evolutionary rate. If branch length are years, then the rates would be in units 1/year.