model (SYM+G) in BEAST2

195 views
Skip to first unread message

Sama

unread,
Feb 21, 2018, 4:57:10 PM2/21/18
to beast-users
Hello,

I'd like to apply SYM+G model to my data in BEAST2, it is not in the models' list, really not sure how to do so, can anyone help me, please

Remco Bouckaert

unread,
Feb 21, 2018, 4:58:34 PM2/21/18
to beast...@googlegroups.com
Hi Sama,

Have a look a the substmodels package, which contains SYM:
https://github.com/BEAST2-Dev/substmodels

Cheers,

Remco

On 22/02/2018, at 10:08 AM, Sama <Sreh...@gmail.com> wrote:

Hello,

I'd like to apply SYM+G model to my data in BEAST2, it is not in the models' list, really not sure how to do so, can anyone help me, please

--
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.

Daria1993

unread,
Mar 17, 2019, 1:04:27 PM3/17/19
to beast-users
Dear Remco, 

I am the beast begginer and I saw you recommend using SSM instead of editing the xml file - even after reading the manual with explanation how to edit xml I'm not sure how to use it. This is my result of jmodeltest2:

 Model selected: 
   Model = SYM+G
   partition = 012345
   -lnL = 3004.3557
   K = 114
   R(a) [AC] =  15.2867
   R(b) [AG] =  81.2855
   R(c) [AT] =  30.8076
   R(d) [CG] =  19.1121
   R(e) [CT] =  116.3970
   R(f) [GT] =  1.0000
   gamma shape = 0.3870 

Do you think that it will be ok for sure if I use SSM package? 

Cheers, 
Daria

Remco Bouckaert

unread,
Mar 18, 2019, 3:13:20 AM3/18/19
to beast...@googlegroups.com
Hi Daria,

You can use the SSM package, and in BEAUti select Standard_SYM as your substitution model. But since SYM is equal to the GTR model with equal frequencies, you could also select GTR in BEAUti and select “All equal” for frequencies.

However, if the site model is of no interest for the question you want to answer, you might want to consider using Bayesian model averaging over all site models that jmodeltest2 tests for by using the bModelTest package (see https://github.com/BEAST2-Dev/bModelTest/wiki for more info). 

Cheers,

Remco
Reply all
Reply to author
Forward
0 new messages