Issue when selecting Birth Death Model prior and tip dates

Ophélie Lebrasseur

unread,

May 15, 2022, 3:22:52 PM5/15/22

to beast-users

Hello all,

I would like to conduct a BETS analyse using Nested Sampling on a dataset with tip dates and using the Birth Death Model prior.

However, upon selecting the setting in Beauti2, a red cross appears next to the Birth Death Model prior stating 'This tree prior cannot handle dated tips. Choose another tip prior'.

I don't quite understand this error, and I've looked everywhere online to try and understand my mistakes but have yet to find a solution. I did come across this possible issue https://github.com/CompEvol/beast2/issues/727 but I still don't understand.

Would anyone be able to help explain where I am going wrong?

Thank you!

Remco Bouckaert

unread,

May 23, 2022, 5:34:39 PM5/23/22

to beast...@googlegroups.com

Hi Ophelie,

The birth-death model assumes all tips are sampled at the same time. If you specify dates tips, BEAUti will warn you that this assumption is broken. BEAST will still be able to run the XML with this model and dated tips, but you cannot rely on it being a birth-death process.

If you have dated tips, you should use a birth-death-sampling model, like BDSKY, which does properly take the tip sampling process in account. Alternatively, coalescent models can handle dated tips, but in this case may be less appropriate for your data.

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 view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/d71e7567-7d4c-42e1-a662-d530066b0c44n%40googlegroups.com.

Ophélie Lebrasseur

unread,

May 24, 2022, 2:27:03 PM5/24/22

to beast-users

Dear Remco,

Thank you so much for your reply, this is very helpful. I had started looking in the published literature as well to see what models researchers working on similar data were using, and the Birth-Death skyline as well as Coalescent models had indeed come up. But I couldn't understand why, so thank you so much for the explanation :)