Fixing a gene tree within a species tree analysis

25 views
Skip to first unread message

random...@gmail.com

unread,
Nov 15, 2018, 12:41:48 PM11/15/18
to beast-users
To the *BEAST experts,

While doing a species tree analysis in *BEAST, one of my gene trees (for mitochondrial DNA) consistently got estimates for the MRCA that were too old (almost triple what the MRCA is as estimated by two independent estimates). My strategy was to fix the mitochondrial tree for the entire run and let the species tree and other gene trees start with random trees. I know this is unusual, but I don't believe the *BEAST estimate for the mitochondrial tree's node ages.
By removing operators in the XML file, I've managed to keep the mitochondrial tree constant across runs. However, I'm not confident that the mitochondrial tree is influencing the species tree estimation now. Does anyone have advice on this?

Best,
Tom

Alexei Drummond

unread,
Nov 15, 2018, 3:21:36 PM11/15/18
to beast...@googlegroups.com
Dear Tom,

How are you calibrating the *BEAST analysis and how have the previous “independent analyses” been calibrated?

If previous analyses have not distinguished between gene tree and species tree, and the fossil calibration used is the same in those analyses and the *BEAST one, but applied to the species tree rather than the gene tree in your analysis then there is every reason for the age of the mtDNA root to be older in your analysis since the mtDNA root must be older than the species tree split.

Alexei
--
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.

random...@gmail.com

unread,
Nov 27, 2018, 1:52:16 PM11/27/18
to beast-users
Alexei,

I think your question highlighted my error. Thank you! In the previous study, I believe the molecular clocks were NOT estimated from a species tree analysis, but were instead estimated each independently given a fixed topology.

-Tom

Alexei Drummond

unread,
Nov 27, 2018, 3:58:46 PM11/27/18
to 'Pedro Bond Schwartsburd' via beast-users
Great :)

Reply all
Reply to author
Forward
0 new messages