search

matsengrp / ecgtheow

Ancestral lineage reconstruction using BEAST or RevBayes
2 stars 2 forks source link

Be able to run different data sets, heh #3

Closed matsen closed 7 years ago

matsen commented 7 years ago

Right now we have a single BEAST .xml file, at 

/home/matsengrp/working/matsen/ecgtheow/runs/2017-07-10/BF520.1-h-IgH.family_0.healthy.tre.seedpruned.100.ids.xml

This clearly needs to be made more flexible. Although we could struggle along with hand-coded xml generation, the right thing is to use a templating language. I very much suggest Jinja2.

Chris Warth has set up a much more sophisticated example than what we have.

I'd take a look at his stuff for inspiration, but build up from minimal examples rather than try to reduce from what he has there.

Please commit templates and code, but again not data.

dunleavy005 commented 7 years ago

May I close this issue? (b/c of https://github.com/matsengrp/ecgtheow/commit/fafbeb7ca7957bb1bb03156eeda86117f69fc1ea)

matsen commented 7 years ago

Have you run it on different data sets? If so, def.

dunleavy005 commented 7 years ago

I re-did BEAST on the same dataset (i.e. use the templating script to auto-generate the BEAST xml), and the plots looked almost the same, nothing new however. I think the other open issue explains how to make a new dataset?

matsen commented 7 years ago

Great. Close.

matsen commented 7 years ago

Oh, actually, do you have the original xml file? Did you do a diff between that and your generated one?

dunleavy005 commented 7 years ago

Yes, they match, outside of the random permutation of sequences/taxon names.

matsen commented 7 years ago

@dunleavy005 can you put some instructions about using your beast templating script into the README? Shoulda asked for this first.

In fact, we could move all the instructions in #4 to the README, including doing a BEAST run. That would be ✨ .

dunleavy005 commented 7 years ago

There is a --help command + the example command run in the README, do you want me to discuss the --help in the README?

But yes, I plan on adding the pre-processing stuff from #4 eventually, once I get some runs in.

matsen commented 7 years ago

😬 heh, somehow I didn't see it. Thanks!