GTR+Gamma model for amino acid data

[Xavi Grau]

Hi everyone,

I've read in the MBE paper that Exabayes implements a GTR model for protein data, besides the fixed-rate matrices (http://mbe.oxfordjournals.org/content/31/10/2553.full):

For protein data, we offer 18 commonly used fixed-rate substitution matrices as well as a GTR model. We also implemented a comprehensive set of topological proposals that assure rapid convergence and adapted these for massively parallel execution.

Yet, it does not appear in the Amino Acid Model Prior section of the manual:

• disc( m 1 = w 1 , m 2 = w 2 , . . ., m n = w n )

a discrete probability distribution assigning weights w i to protein substitution

matrices m i . If only one model is specified, this is equivalent to a fixed prior.

m may be one of the following models: DAYHOFF, DCMUT, JTT, MTREV,

WAG, RTREV, CPREV, VT, BLOSUM62, MTMAM, LG, MTART, MTZOA,

PMB, HIVB, HIVW, JTTDCMUT, FLU. 

Therefore, if I understand it correctly, is GTR activated by default if I don't specifiy a fixed-rate protein model as a prior? And, also, I assume the same is true for the activation of the gamma distribution of rate heterogeneity, as you've said before for nucleotide data (https://groups.google.com/forum/#!searchin/exabayes/gtr/exabayes/yHJ93loP3jg/etjNiHD6xuwJ)

Thanks in advance, and excuse me if this is a silly question. I just couldn't find a definite answer in the docs.

Best,

Xavi

[Andre Aberer]

Hi Xavi,

I am sorry, the manual is not 100% up to date. But the default config
file is.

Yes, the Gamma model for rate heterogeneity is always on by default and
cannot be turned off currently.

You can turn on GTR for AA by setting the substitution rate prior
("revMatPr") to "dirichlet" (i.e., remove the comment around that
awkwardly long statement). If all 189 values are left at 1.0, then you
have the equivalent of a multi-dimensional uniform prior.

Maybe you'll also have to comment out the "state frequency prior"
(statefreqpr) and set it to dirichlet.

As a side note, if you want to use fixed-rate matrices but have ExaBayes
sample the state frequencies, you only have to set statefreqpr to
dirichlet.

--
Best regards,
Andre

Xavi Grau writes:

> Hi everyone,
>
> I've read in the MBE paper that Exabayes implements a GTR model for protein
> data, besides the fixed-rate matrices
> (http://mbe.oxfordjournals.org/content/31/10/2553.full):
>

> *For protein data, we offer 18 commonly used fixed-rate substitution

> matrices as well as a GTR model. We also implemented a comprehensive set of
> topological proposals that assure rapid convergence and adapted these for

> massively parallel execution.*

>
>
> Yet, it does not appear in the Amino Acid Model Prior section of the manual:
>
>

> *• disc( m 1 = w 1 , m 2 = w 2 , . . ., m n = w n )*
> *a discrete probability distribution assigning weights w i to protein
> substitution*
> *matrices m i . If only one model is specified, this is equivalent to a
> fixed prior.*
> *m may be one of the following models: DAYHOFF, DCMUT, JTT, MTREV,*
> *WAG, RTREV, CPREV, VT, BLOSUM62, MTMAM, LG, MTART, MTZOA,*
> *PMB, HIVB, HIVW, JTTDCMUT, FLU. *

[Xavi Grau]

Hi Andre,

Thanks a lot, I'll give it a try.

Best,

Xavi

El dia 01/06/2016 8:44 a. m., "Andre Aberer" <an...@aberer.io> va escriure:

--
You received this message because you are subscribed to a topic in the Google Groups "ExaBayes" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/exabayes/_NJzyBVZpxY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to exabayes+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Christian Rinke]

To follow up on this older post - what is actually the default amino acid substitution model in ExaBayes? Assuming we don't specify any particular model using -m?

Best,

Chris

[Alexandros Stamatakis]

I am not sure actually, I'd assume it integrates over all possible AA
models using a uniform prior, please have a look at the two examples:

stadard models:
https://github.com/aberer/exabayes/blob/devel/examples/aa/config.nex

and

GTR model:
https://github.com/aberer/exabayes/blob/devel/examples/aa-revmat/config.nex

hope that helps,

Alexis

On 31.07.19 03:52, Christian Rinke wrote:
> To follow up on this older post - what is actually the default amino
> acid substitution model in ExaBayes? Assuming we don't specify any
> particular model using -m?
> Best,
> Chris
>
>
> On Wednesday, 1 June 2016 02:02:25 UTC+10, Xavi Grau wrote:
>
> Hi everyone,
>
> I've read in the MBE paper that Exabayes implements a GTR model for
> protein data, besides the fixed-rate matrices
> (http://mbe.oxfordjournals.org/content/31/10/2553.full

> <http://mbe.oxfordjournals.org/content/31/10/2553.full>):
>
> /For protein data, we offer 18 commonly used fixed-rate

> substitution matrices as well as a GTR model. We also
> implemented a comprehensive set of topological proposals that
> assure rapid convergence and adapted these for massively

> parallel execution./

>
>
> Yet, it does not appear in the Amino Acid Model Prior section of the
> manual:
>

> /
> /
> /• disc( m 1 = w 1 , m 2 = w 2 , . . ., m n = w n )/
> /a discrete probability distribution assigning weights w i to
> protein substitution/
> /matrices m i . If only one model is specified, this is
> equivalent to a fixed prior./
> /m may be one of the following models: *DAYHOFF, DCMUT, JTT,
> MTREV,*/
> /*WAG, RTREV, CPREV, VT, BLOSUM62, MTMAM, LG, MTART, MTZOA,*/
> /*PMB, HIVB, HIVW, JTTDCMUT, FLU. */

>
>
>
> Therefore, if I understand it correctly, is GTR activated by default
> if I don't specifiy a fixed-rate protein model as a prior? And,
> also, I assume the same is true for the activation of the gamma
> distribution of rate heterogeneity, as you've said before for
> nucleotide data
> (https://groups.google.com/forum/#!searchin/exabayes/gtr/exabayes/yHJ93loP3jg/etjNiHD6xuwJ

> <https://groups.google.com/forum/#!searchin/exabayes/gtr/exabayes/yHJ93loP3jg/etjNiHD6xuwJ>)

>
> Thanks in advance, and excuse me if this is a silly question. I just
> couldn't find a definite answer in the docs.
>
> Best,
>
> Xavi
>

> --
> You received this message because you are subscribed to the Google
> Groups "ExaBayes" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to exabayes+u...@googlegroups.com
> <mailto:exabayes+u...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/exabayes/97261777-7bfa-4df3-be00-5270c06d2a2b%40googlegroups.com
> <https://groups.google.com/d/msgid/exabayes/97261777-7bfa-4df3-be00-5270c06d2a2b%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology

www.exelixis-lab.org

[Andre Aberer]

Alexis is right on both accounts.
-Andre