Treeannotator and Posterior Probability Limit

[Matthías Macé]

Dear Beast Users,
Does anyone found a problem when using Treeannotator and Posterior Probability Limit (PPL) box ?
Indeed, when I am trying to summarise trees from MrBayes with a PPL of 0.5 (and above) and when I am trying to open the result in FigTree, it always shows nodes with PPL below.
I am using Treeannotator on macosx 10.5 (but it is the same with 10.4).
Is there any way to fix this or did I misunderstand this function ?
Thanks a lot for the help,
Beast Regards, ;-)
MM

--
Dr Matthias Macé, Vet. Med. Post-Doctoral Fellow / Chargé de Recherches Cliniques CHU

!! Attention, nouvelle adresse !! !! Take care of the new address !!

INSERM U563, Centre de Physiopathologie de Toulouse Purpan

Département Génétique

Génétique des troubles de la réfraction et des épithéliums oculaires-Eq. F Malecaze/P Calvas

C.H.U. Purpan - BP 3028

31024 TOULOUSE CEDEX 3

tel: (33) 562 74 45 09

mobile: (33) 608 881481

http://www.u563.toulouse.inserm.fr/

[Andrew Rambaut]

Dear Matthias,

TreeAnnotator doesn't produce consensus trees. It annotates a given
tree with information from the posterior sample. Its default is to
find a tree in the sample with the Maximum Clade Credibility and use
that. The PP Limit value sets the minimum PP for which it annotates a
node (with height HPDs, rate, etc.). This is because if there are only
a few trees with that clade present in the sample these estimates will
be very poor.

I would be cautious using TreeAnnotator on MrBayes trees unless they
are rooted (i.e., a molecular clock model) as the program calculates
clade frequencies not split frequencies so may produce a misleading
result on unrooted trees.

Andrew

On 23 Feb 2008, at 17:16, Matthías Macé wrote:

> Dear Beast Users,
> Does anyone found a problem when using Treeannotator and Posterior
> Probability Limit (PPL) box ?
> Indeed, when I am trying to summarise trees from MrBayes with a PPL
> of 0.5 (and above) and when I am trying to open the result in
> FigTree, it always shows nodes with PPL below.
> I am using Treeannotator on macosx 10.5 (but it is the same with
> 10.4).
> Is there any way to fix this or did I misunderstand this function ?
> Thanks a lot for the help,
> Beast Regards, ;-)
> MM
>
> --
> Dr Matthias Macé, Vet. Med. Post-Doctoral Fellow / Chargé de

> Recherches Cliniques CHU!! Attention, nouvelle adresse !! !! Take
> care of the new address !!INSERM U563, Centre de Physiopathologie de
> Toulouse PurpanDépartement GénétiqueGénétique des troubles de la

> réfraction et des épithéliums oculaires-Eq. F Malecaze/P
> CalvasC.H.U. Purpan - BP 302831024 TOULOUSE CEDEX 3tel: (33) 562 74
> 45 09mobile: (33) 608 881481http://www.u563.toulouse.inserm.fr/
>
> >

___________________________________________________________________
Andrew Rambaut
Institute of Evolutionary Biology University of Edinburgh
Ashworth Laboratories Edinburgh EH9 3JT
EMAIL - a.ra...@ed.ac.uk TEL - +44 131 6508624

[Jake Gregg]

Dr. Rambaut,

I am having a problem similar to Mathias above.  However, with trees generated in BEAST.  I expect nodes that are under 0.95 posterior prob to be collapsed when I open the .tre file in figtree, but the topology and statistics for these nodes are still present.  I guess I don't understand what setting the postprob limit to 0.95 in treeannotator really does.

Any help you can offer would be appreciated.

Jake Gregg

USGS Marrowstone Marine Field Station

[Lambodhar Damodaran]

Hey Jake,

When you use treeanotator it does not prune the clades that are below the posterior prob set, it simply does not annotate any node that does not have 95% posterior probability or greater. In figtree I would try looking at the node label and node bar settings. This is useful if you want to identify the clades that are well supported. You can collapse the clades in figtree manually, or doing it through code using a package like ggtree in R. Hope this is helpful!

Lambodhar

[Jake Gregg]

Thanks,

I'll check out ggtree.

Jake

--
You received this message because you are subscribed to a topic in the Google Groups "beast-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beast-users/i0lQrLHSBmo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to beast-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/f23eda65-25e2-4c1f-b2ed-b4283185afe7n%40googlegroups.com.