ModelTest-NG – zero protein models

[Sergios-Orestis Kolokotronis]

Hello, is this the right group for ModelTest-NG. I see mentions of it both here and in the jModelTest group.

Compiling it for OSX is a bit problematic right now, so I installed the previous version 0.1.1 GUI. I load a Fasta AA alignment and a tree, set 10 threads, and in Settings the RAxML template (and MT and others) and even selected only a few or just one model with uniform and Gamma. I press Run and I get an error pop-up window saying there are 0 models to optimize. Somehow it doesn't recognize the already selected boxes for models in Settings. Also tried without a tree or a MP tree.

Basically, the console does not reflect the options I selected in Settings.

Thanks in advance for your response.

Best,

Sergios

Loaded alignment

Format: FASTA

Datatype: Protein

Num.Sequences: 27

Sequence Length: 251

Site patterns: 199

Building modeltest instance for 10 threads

--------------------------------------------------------------------------------

Input data:

MSA: /path/to/my/aln

Tree: Maximum likelihood

file: -

#taxa: 27

#sites: 251

#patterns: 199

Output:

Log:

Results:

Selection options:

# protein matrices: 21

# protein models: 0

include model parameters:

Uniform: false

p-inv (+I): false

gamma (+G): false

both (+I+G): false

fixed freqs: false

estimated freqs: false

#categories: 4

asc bias: none

epsilon (opt): 0.01

epsilon (par): 0.001

Additional options:

verbosity: low

threads: 10/12

RNG seed: 12345

subtree repeats: disabled

--------------------------------------------------------------------------------

/path/to/my/tree

Loaded tree

Building modeltest instance for 10 threads

--------------------------------------------------------------------------------

Input data:

MSA: /path/to/my/aln

Tree: Fixed user

file: /path/to/my/tree

#taxa: 27

#sites: 251

#patterns: 199

Output:

Log:

Results:

Selection options:

# protein matrices: 21

# protein models: 0

include model parameters:

Uniform: false

p-inv (+I): false

gamma (+G): false

both (+I+G): false

fixed freqs: false

estimated freqs: false

#categories: 4

asc bias: none

epsilon (opt): 0.01

epsilon (par): 0.001

Additional options:

verbosity: low

threads: 10/12

RNG seed: 12345

subtree repeats: disabled

--------------------------------------------------------------------------------

Building modeltest instance for 10 threads

--------------------------------------------------------------------------------

Input data:

MSA: /Volumes/mammut2/Dropbox/RESEARCH/Borrelia/data/CspA/CspA_aa_sml.mcof.fasta

Tree: Fixed user

file: /Volumes/mammut2/Dropbox/RESEARCH/Borrelia/data/CspA/CspA_aa_sml.mcof.T3.raxml.bestTree

#taxa: 27

#sites: 251

#patterns: 199

Output:

Log:

Results:

Selection options:

# protein matrices: 21

# protein models: 0

include model parameters:

Uniform: false

p-inv (+I): false

gamma (+G): false

both (+I+G): false

fixed freqs: false

estimated freqs: false

#categories: 4

asc bias: none

epsilon (opt): 0.01

epsilon (par): 0.001

Additional options:

verbosity: low

threads: 10/12

RNG seed: 12345

subtree repeats: disabled

--------------------------------------------------------------------------------

[Alexandros Stamatakis]

Hola Diego,

Where should we provide support for Modeltest-NG?

I guess it's better to do it via your google group?

Thanks,

Alexis

On 21.12.2017 02:04, Sergios-Orestis Kolokotronis wrote:
> Hello, is this the right group for *ModelTest-NG*. I see mentions of it

> --
> You received this message because you are subscribed to the Google
> Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to raxml+un...@googlegroups.com
> <mailto:raxml+un...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

--
Alexandros (Alexis) Stamatakis

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

www.exelixis-lab.org

[Sergios-Orestis Kolokotronis]

Which group is that? I'm not seeing a Modeltest-NG group and the jModeltest group doesn't seem oriented towards NG (at least after a couple of searches). Maybe I missed something? Thanks in advance–

[Alexandros Stamatakis]

Hola Diego,

Can you help?

And happy new year to everyone of course :-)

Alexis

On 06.01.2018 07:25, Sergios-Orestis Kolokotronis wrote:
> Which group is that? I'm not seeing a *Modeltest-NG* group and the

> > an email to raxml+un...@googlegroups.com <javascript:>
> > <mailto:raxml+un...@googlegroups.com <javascript:>>.

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.

>
> --
> Alexandros (Alexis) Stamatakis
>
> Research Group Leader, Heidelberg Institute for Theoretical Studies
> Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
>

> www.exelixis-lab.org <http://www.exelixis-lab.org>

[Diego Darriba]

Hi Sergios,

Sorry for the huge delay. There are 2 bugs here... LG4X/M models are not addressed correctly by the
GUI, and the error message is wrong. The actual problem is not that there are 0 protein models, but
the model count differs internally between the expected and the observed number of models. I fixed
both already, but I'll need a few more days to find a mac to build and test the OS X binaries. Thank
you for your report!

BTW, I also think it will be better to use jModelTest group for questions regarding modeltest-ng.

Best,
Diego.