can geographic traits be specified via BEAUti yet?

[Rachel]

it seems as though folks are able to use the new build v1.7.2 to perform discrete phylogeographic analyses, but are you folks achieving this via xml editing or via BEAUti?
I have so many taxa that I *KNOW* errors will be made if I try to edit the xml. It seems tantalizingly close in BEAUti, but I cannot find any specification for how geographic data is to be collated and entered via BEAUti.
thanks in advance!
Rachel

[Philippe Lemey]

Dear Rachel,

You can indeed set up discrete diffusion models in BEAUti 1.7.2 and I would recommend to do so because the xml specification has evolved considerably since the tutorials were put together. In addition to the standard model setup in BEAUti, you need to go to the Traits panel and add a trait. You can import traits from a mapping file;  for an example of this, please see the taxaLocations.txt file in the examples/discreteTraits/ folder in BEAST 1.7.2. that goes with the data in the H5N1_HA.nex file. A trait should appear in the left window. Click on that and the taxa with corresponding traits will appear in the window on the right. Then click "Create partition from Trait". Under the Sites panel, you should now see the partition under "Substitution Model" on the left. Click on it to select the discrete trait Substitution model and proceed with the prior specification for the discrete trait diffusion process and you should be ready to go if you have set all the other models as well.

Kind regards,

Philippe

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To view this discussion on the web visit https://groups.google.com/d/msg/beast-users/-/4myjZCUAm0YJ.
To post to this group, send email to beast...@googlegroups.com.
To unsubscribe from this group, send email to beast-users...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/beast-users?hl=en.

[Rachel L]

Philippe-

thank you for your assistance- I can't wait to give it a try!

Rachel

[Rachel Life]

Philippe-

I'm playing with setting up my discrete run now, and I have a question- can I use latitude and longitude for my locations instead of place names? I've been able to get them accepted in the traits tab of BEAUti. Is the post-run analysis (ie making the kml file) the same in this version of BEAST? 

Rachel

[Jacob Berv]

Ditto - Is it possible to perform continuous diffusion models using lat/long as well? What is the appropriate format for the mapping file?

[Andrew Rambaut]

Dear Jacob,

It is indeed possible. Simply put the lat/long values (decimal degrees) in a tab delimited file with the taxon label in the first column. Import as traits and then select both traits and click 'Create partition from traits'. You can then select a diffusion model on them. An example alignment and lat/long file is in 'examples/continuousTraits' folder (North American Rabies named RABV*).

The reconstruction can then be done using the application 'SPREAD' in addition to the normal post-BEAST diagnostics.

http://www.phylogeography.org/SPREAD

Andrew

On 17 Aug 2012, at 23:40, Jacob Berv <jacob...@gmail.com> wrote:

> Ditto - Is it possible to perform continuous diffusion models using lat/long as well? What is the appropriate format for the mapping file?
>

> --
> You received this message because you are subscribed to the Google Groups "beast-users" group.

> To view this discussion on the web visit https://groups.google.com/d/msg/beast-users/-/0O63Ralo-F0J.

> To post to this group, send email to beast...@googlegroups.com.
> To unsubscribe from this group, send email to beast-users...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/beast-users?hl=en.

___________________________________________________________________
Andrew Rambaut
Institute of Evolutionary Biology University of Edinburgh
Ashworth Laboratories Edinburgh EH9 3JT
EMAIL - a.ra...@ed.ac.uk TEL - +44 131 6508624


--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

[Gehara]

Dear Andrew,

I've been performing the continuous phylogheographic analysis with different datasets. When I generate the XML file via Beauti something strange happens in the MCMC. The chain starts without any problem, but at some point the prior, posterior, and likelihood values are shown as "?" and the traces of all parameters are interrupted or shown as a straight line. I've tried different datasets and this always happens. However if I build the XML myself according to the tutorial available in the Beast website everything works smoothly, both Homogeneous Brownian Diffusion and Relaxed Random Walk.
I've noticed that the XML file generated by Beauti does not contain the jitter code, I wonder if this might be the problem given that I have equal traits for some samples.
Anyway, I already got nice results with the analysis, I'm just reporting the issue I had.

Best regards

[Andrew Rambaut]

The issue is if you have multiple strains with exactly the same geographic coordinates. Under the model, this should never happen but obviously sampling location is not always that precise. The jitter code adds some random noise so that the model actually runs but what effect this has on the results depends on how much jitter you add (ideally you want to add something related to the precision of the sampling).

At the moment, BEAUti doesn't add the jittering option.

Andrew