Make BEAUti2 more efficient for multi-locus data

[Santiago Sánchez]

Dear developers,

This is just a plea to make future versions of BEAUti2 more efficient and practical for multi-locus data (e.g. in a similar way, perhaps, than in BEAUti1). For people working with multi-locus data (I include myself) it becomes very cumbersome and error prone to specify the same priors and settings over and over for each partition.

For instance, selecting the EBSP tree prior for every partition is a waste of time (in BEAUti1 this only done once). It also would be nice to be able to "clone" priors, for instance use the same prior distribution for all clock rates, with a few clicks. Similarly, in the MCMC tab, I need to select a "store every" number for the MCMC, which doesn't translate to other logfiles (in BEAUti1 this only done once), not even the "storeEvery" number in the state file that is being saved for restarting (this has to be done manually by opening the XML). So here again, many, many clicks, which can lead to error by repeating the same thing over and over.

I think that BEAST2 is really cool, and I love the restart functionality, but I typically don't use it because of these issues (its just very time consuming to generate XML files for multi-locus data). The last time it took me about 15/20 minutes to prepare an XML for 20 loci. In contrast, the same analysis took me less that 3 minutes to prepare in BEAUti1. Imagine if I had data for 100 loci or more!

Sincerely,

Santiago

--

==========================

Santiago Sanchez-Ramirez, PhD

Postdoctoral Associate

Ecology and Evolutionary Biology

University of Toronto

==========================

[Remco Bouckaert]

Hi Santiago,

Thanks for the feedback, and I raised an issue (#673) to address this.

There is a way to clone clock priors though. It is probably easiest to link clock models (in the partition panel), then set up the clock model and prior for the single clock model (in the priors panel), then unlink clock models (in the partition panel), and at that point all clock models and their priors will be cloned.

I will have a look at the other problems.

Cheers,

Remco

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beast-users...@googlegroups.com.
To post to this group, send email to beast...@googlegroups.com.
Visit this group at https://groups.google.com/group/beast-users.
For more options, visit https://groups.google.com/d/optout.

[Santiago Sánchez]

Hi Remco,

Thanks a lot!! Now that you mentioned it, I recall you have suggested this way of cloning clock priors before; I simply forgot about it :-).

Hopefully these changes will make BEAST2 more appealing to multi-locus analysts!

Best,

Santiago