Bayesian Skygrid settings :(

[Artem B]

Dear all,
I try to figure out what the "grid-point" value I should set and how it affects on my analysis.
I know that 'time at last point' should be set with the expected root height but I can't understand what the 'grid-point' value will be correct.

BEAST version 1.8.4

Regards

[Nahog Aa]

Hi,

Did you resolve this?

My root height ranges from 0-30. I am not sure if the "time at last point" should be set as 1.0 as in the tutorial papers.

Many thanks.

[Pavel Rinkman]

Hello!

You can get a good estimate through linear regression on ML tree, for instance, through TempEst & set tMRCA as time at last transition point.

среда, 3 августа 2022 г. в 21:47:32 UTC+3, gaode...@gmail.com:

[Artem B]

Hello!
A random seed may give an initial root height value higher than 1.0 and, therefore, your analysis will not even start (an error log will say the same).
Try to set the last point value more than 30 (e.g., 35). As far as I remember, 35 or 350 doesn't affect analysis results.

Best wishes,
Artem

четверг, 4 августа 2022 г. в 02:47:32 UTC+8, gaode...@gmail.com:

[Artem B]

Yeah, that's really good idea! Thank you.

Best wishes,
Artem

четверг, 4 августа 2022 г. в 06:13:22 UTC+8, Pavel Rinkman:

[Pavel Rinkman]

As far as I remember, 35 or 350 doesn’t affect analysis results.

With predefined number of parameters it sets the scale of grid as

So, in my view, this is a big difference

​

четверг, 4 августа 2022 г. в 05:15:28 UTC+3, ui.ar...@gmail.com:

[Artem B]

I reread Gill et al 2013 and it seems you are right.
It is the best to run analysis with different cut off values and compare results, I guess.

Cheers,
Artem

четверг, 4 августа 2022 г. в 15:35:34 UTC+8, Pavel Rinkman:

[Pavel Rinkman]

My sentiments exactly! Although I maintain frequency via controlling points number. Because my object has a history spanning about a couple of centuries but most of the samples has been being taken during thirty years with obvious change in population size over the period.

Best wishes,

Pavel

пятница, 5 августа 2022 г. в 05:02:58 UTC+3, ui.ar...@gmail.com:

[Nahog Aa]

Many thanks! I set it as 30 and it worked!

Best,

Han

Artem B <ui.ar...@gmail.com> 于2022年8月4日周四 10:15写道：

--
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/10a869bf-774b-43dd-b2c2-4b3e934bc817n%40googlegroups.com.