Explicitly reporting observed amino acid frequencies of the alignment

[Joran Martijn]

Hi!

I'm currently trying to run some parametric simulations using SiteSpecific.seq-gen (http://www.mathstat.dal.ca/~hcwang/Procov/SiteSpecific.seq-gen/). Ideally, I would like to run simulations with the +F version of the model, meaning I need to extract the observed amino acid frequencies from the alignment, and feed them into the simulation software.

It would therefore be great if IQTREE would report the observed frequencies it uses for the +F somewhere (maybe in the .log, or a separate output file?).

I could write a little script myself, but since I want to be absolutely sure I'm using the same +F frequencies as IQTREE, I feel this would be a bit better.

My apologies if IQTREE already does this, but I wasn't able to find them.

Cheers,

Joran

[Bui Quang Minh]

Hi Joran,

When you run the protein analysis with -m LG+F for example, then the amino-acid frequencies will be printed into .iqtree file like this:

SUBSTITUTION PROCESS

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

Model of substitution: LG+F

State frequencies: (empirical counts from alignment)

  pi(A) = 0.0790

  pi(R) = 0.0395

  pi(N) = 0.0358

  pi(D) = 0.0544

  pi(C) = 0.0193

  pi(Q) = 0.0487

  pi(E) = 0.0513

  pi(G) = 0.0881

  pi(H) = 0.0245

  pi(I) = 0.0509

  pi(L) = 0.0932

  pi(K) = 0.0741

  pi(M) = 0.0237

  pi(F) = 0.0433

  pi(P) = 0.0423

  pi(S) = 0.0604

  pi(T) = 0.0561

  pi(W) = 0.0136

  pi(Y) = 0.0372

  pi(V) = 0.0646

Which you can use. Moreover, I CC my collaborator, Huaichun here, in case he has further tip.

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.

--

Bui Quang Minh

Center for Integrative Bioinformatics Vienna (CIBIV)

Campus Vienna Biocenter 5, VBC5, Ebene 1

A-1030 Vienna, Austria

Phone: ++43 1 4277 74326

Email: minh.bui (AT) univie.ac.at

[Joran Martijn]

Hi Minh,

I can't find this information in the .iqtree file. Under substitution process, it only reports the mixture weights for the 60 components and the F component, and the gamma relative rate parameters. Maybe it doesnt report the observed frequencies when you invoke a mixture model?

For this analysis, I'm using version 1.5.0, and invoked the model LG+C60+F+G.

For now I'll just run a dummy LG+F analysis to get the frequencies, but I suppose it would be nice to have them reported when you run LG+C60+F+G or similar models as well.

Joran

[Heiko Schmidt]

Dear Joran,

I am not completely sure what you need. If you need the observed column-wise AA frequencies (a wild guess because you use SiteSpecific.seq-gen), I think this is not output by IQ-Tree.

If I remember correctly, I had implemented that output into TREE-PUZZLE sometime in the past.
If that is what you need… I can dig up the command line options required to get this. However, TREE-PUZZLE requires standard PHYLIP/Newick input for the alignment, that means the whole alignment has to be in one file and sequence names are only allowed up to 10 letters. For the latter restriction (10 letter) there is a pretty easy way to ensure this, because just for extracting the site AA frequencies, you do not need to keep the original names. (I can provide more information if required.)

Best,
Heiko

-----------------------------------------------------------------------------
Heiko Schmidt

Center for Integrative Bioinformatics Vienna (CIBIV)

University of Vienna / Max F. Perutz Laboratories (MFPL)
Campus Vienna Biocenter 5 (VBC5)
A-1030 Vienna, Austria
http://www.cibiv.at/
-----------------------------------------------------------------------------

[Joran Martijn]

Hej Heiko,

IQ-TREE does output the overall observed amino acid frequencies, which is what I need. It just doesn't seem to output it when I specify a mixture model. It did output them when using LG+F , like Minh said.
And yes, the frequencies are meant for SiteSpecific.seq-gen. Huaichun updated the tool recently so you can input your own amino acid frequencies.

So, for me the problem is solved. I just run IQ-TREE with LG+F to get the observed frequencies of a dataset, and then use those as input for SiteSpecific.seq-gen. I was just pointing out that it was a bit strange that if you specify a mixture model in addition to +F, IQTREE does not output the observed overall amino acid frequencies.

Cheers,

Joran

[Bui Quang Minh]

Hi Joran again, I put this into the TODO list…

M