How to select model for binary irreversible sequence data?

[Chang Ye]

mix package in phylip is for building tree from binary irreversible sequence.

Changes 0 --> 1 are much more probable than changes 1 --> 0 (Camin-Sokal)

Seems that there is no model in iqtree to do this job. Which argument should I modify to make iqtree fit Camin-Sokal?

• http://evolution.genetics.washington.edu/phylip/doc/mix.html

​

[Chang Ye]

Is it possible to assign prior mutation rate for any kind of mutation?

Just like GTR{1.0,2.0,1.5,3.7,2.8} but include more number in the blacket, XXX{1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0}

[Minh Bui]

Hi there, I’ve never heard of Camin Sokal model. The web page you sent talks about parsimony, whereas IQ-TREE does likelihood inference. So I don’t know what is this model in the likelihood context…

Cheers

Minh

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

[Minh Bui]

Hi again, 

I’m assuming you want to perform non-reversible DNA model. This most general model is called 12.12, which belongs to the class of Lie Markov models. You can fix Lie Markov model parameters with:

-m “12.12{r1,…,r11}"

However, r_i are parameters in the range (-1,1) and have a very different meaning than GTR model. For more details please have a look at this paper: (Woodhams et al., 2015)

Cheers

Minh

[Chang Ye]

Hi Minh, 

I read document about Lie Markov model carefully, and make some attempt to apply the model to BINARY data. 

But Seems that these models are for DNA only, even if I manually encode the 1, 0  chars into two DNA base, iqtree won't start...

It is possible to run iqtree on Binary data with irreversible model? If I like to use such model with iqtree, what can I do? Could you please give me some suggestions?

sincerely,

Chang

On Monday, October 8, 2018 at 2:39:58 PM UTC+8, Minh Bui wrote:

Hi again, 

I’m assuming you want to perform non-reversible DNA model. This most general model is called 12.12, which belongs to the class of Lie Markov models. You can fix Lie Markov model parameters with:

-m “12.12{r1,…,r11}"

However, r_i are parameters in the range (-1,1) and have a very different meaning than GTR model. For more details please have a look at this paper: (Woodhams et al., 2015)

Cheers

Minh

On 8 Oct 2018, at 5:07 pm, Chang Ye <yech...@gmail.com> wrote:

Is it possible to assign prior mutation rate for any kind of mutation?

Just like GTR{1.0,2.0,1.5,3.7,2.8} but include more number in the blacket, XXX{1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0}

On Sunday, October 7, 2018 at 11:14:12 PM UTC+8, Chang Ye wrote:

mix package in phylip is for building tree from binary irreversible sequence.

Changes 0 --> 1 are much more probable than changes 1 --> 0 (Camin-Sokal)

Seems that there is no model in iqtree to do this job. Which argument should I modify to make iqtree fit Camin-Sokal?

• http://evolution.genetics.washington.edu/phylip/doc/mix.html

​

​

--
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.

To unsubscribe from this group and stop receiving emails from it, send an email to iqt...@googlegroups.com.

[Chang Ye]

Also there is no useful error message for debugging...

```bash

ERROR: STACK TRACE FOR DEBUGGING:

ERROR:

ERROR: *** IQ-TREE CRASHES WITH SIGNAL SEGMENTATION FAULT

```

[Minh Bui]

Hi Chang,

Somebody may correct me. But currently I believe there is no non-reversible Markov process for binary data, unless this process is non-stationary, i.e. the state frequencies will change along the tree. However, non-stationary models are not supported in IQ-TREE and it’ll take a large effort to implement it. Therefore, right now I suggest that you look for other software… 

Cheers,

Minh

To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.

To post to this group, send email to iqt...@googlegroups.com.
Visit this group at https://groups.google.com/group/iqtree.

To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/2cee3f59-30ec-4000-a135-1b019c6589b4%40googlegroups.com.

[Chang Ye]

Hi Minh

Thanks for your advice.