Pipelines for building PhyDyn xmls

[m-d-grunnill]

I have been going through some PhyDyn tutorials online. I was wondering what work there has been on pipeilines for building PhDyn xml files for running on BEAST. I ask as I don't want to reinvent the wheel.

I have previously worked on UKHSA's PyGOM and I think I can see how that could be extended to produce such a pipeline. I was thinking an ODE model could be coded up in a PyGOM model object. The PyGOM could take initial values using current methods. A new method of the PyGOM model would take the following arguments:

• String: Path to fasta file.
• String: Seperator to extract date out of sequence tags.
• String: Date format.
• A list/set of stings: Infected/deme states
• Dictionary: ODE parameter bounds.
• Dictionary: ODE parameter priors.
• Dictionary: Evoluationary/Phylo parameters.
• Dictionary: Evoluationary/Phylo priors.
• String: Save location for ready to run BEAST xml.

I think PyGOM could be modified to this purpose as it already has methods for determining birth and progression matrices as part of determining R0. Such a method could be modified to produce birth, migration (progression) and death matrices that are used in PhyDyn.

[igococha]

Hi Martin,
Thanks for taking an interest in PhyDyn! There´s an R module developed by Erik that I have to review (it's been a few months now, a busy 3 months) that generates xml ready to be read by Beauti for further editing, if needed. I probably have to work on the the backend ie Beauti templates. Let me check PyGOM to see if it could be useful given that some work has been done in the R approach. A great excuse to get back to PhyDyn!

[emvolz]

The R package is here https://github.com/emvolz/phydynbeast
It is useable, but does not have all features implemented.
I think a python package could also be useful.

[m-d-grunnill]

Appologies been away from work for 2 weeks. @emvolz thanks I have installed the phydynbeast R package and I am going to test it a bit.