BEASTX reports TMRCA for a different node than BEAST 1.10.4 and BEAST2

[Kurt Wollenberg]

Hello BEASTX developers:

For BEASTX, BEAST 1.10.4, and BEAST2 I ran 4 independent runs of 20M generations each following the “Rates and Dates” tutorial with the yellow fever virus sequence data.

BEASTX

All runs had ESS between 274 to 374

age(Americas), run 4 (highest ESS) 1323.8911 [1057.5344, 1564.3789]

Looking at the HIPSTR summary tree for this run 1323 is the age of the node connecting the Americas clade with the West African clade. The basal node for the Americas clade has an age of 1635.3962.

 

BEAST v1.10.4

All runs had ESS range of 234 to 368

age(Americas), run 3 (highest ESS) 1603.7158 [1413.8725, 1765.3615]

 

BEAST2

Runs had ESS from 7198 to 7643

age(Americas), run 3 (highest ESS) 1606.078 [1559.0348, 1655.339]

 

BEASTX is returning the age/tmrca estimate for one node deeper in the tree than BEAST 1.10.4 and BEAST2. Perhaps the index telling the program which node to report is off by one?

Cheers,

Kurt Wollenberg, Ph.D. 

Contractor – Guidehouse
Phylogenetics Specialist 
Scientific Support Section 
Bioinformatics and Computational Biosciences Branch (BCBB) 
OCICB/OSMO/OD/NIAID/NIH

[Alexei Drummond]

Hey Kurt,

Strange that the HPD is significantly narrower for BEAST2 run compared to BEAST v1.10.4. Are the priors the same?

Cheers

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 view this discussion visit https://groups.google.com/d/msgid/beast-users/d366a473-9bb7-4ee1-a2a5-662f3d84d9c6n%40googlegroups.com.