mrca age approaching 0

[ABedoya]

Hello all,

I am using starbeast in beast2 to estimated a time-calibrated species tree for >100 species with 15 loci. I added monophyletic constrains, a starting tree, and specified a calibration node with a normal distribution, mean 35.6 Ma, SD 5. Analysis is running but after over 200 million generations, mrca age  is 8.4E-2. I am using a calibrated yule prior. Do you have any ideas for why this is happening and how I can fix this?

[Jordan Douglas]

Hi,

Did you also estimate the clock rate? If you leave it fixed at 1 susbtitution per site per unit of time, then it will conflict with the MRCA prior of 36.6 units of time and if the alignment is quite big then the data (likelihood) will overpower the prior and the height of the tree will be essentially the same as what it would be without the mrca prior.

There is a good guide on selecting clock rate priors here:

https://beast2-dev.github.io/hmc/hmc/StrictClock/clock.rate/

Jordan 

[ana maria Bedoya]

Hello Jordan,

Thank you so much for your quick response! I figured that from past answers of yours to similar questions. I can confirm this fixed my problem.

Best,

Ana

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/7d977ae0-b84a-4d15-8a17-5a66ba806664n%40googlegroups.com.