ancestral sequence reconstruction

[Liang Wang]

Hi, everyone:

    I am a newcomer in reconstruction of ancestral sequences. There are 2 question about using bppsuite to reconstruction of ancestral sequences:

    1. As the best substitution model of our data is FLU+I+G which was determined by prottest was not implicated in bppsuite, I was wondering if is was possible to use user defined substitution model in bppml? If it was possible, is ther some tutorial of exmple to illustrate how to use it? If it was impossible, was there any alternatives to solve this problem?

    2. As I have not been used bppancestor before, is there any template of control file or examples of bppancestor which I could learned and then modified according to our data?

[Laurent Guéguen]

Dear Liang,

there are many models in bppsuite, but it is generally not possible to describe directly its own model without doing it in bpp-phyl library directly.

However, some features of a model, such as equilibrium distribution, can be described directly. So perhaps you could describe your model

more precisely (or give me a biblio ref), so that we can see what you could do?

Otherwise, I can easily implement it in bpp-phyl, usually.

About the configuration file, in bppsuite there are several examples of in Examples/MaximumLikelihood directories, see the .bpp files.

In these files, there is all the syntax to describe data, models, etc, for all bppsuite programs. Actually, the specific options for

bppancestor are not used there, but they can be added, as they are described in:

http://biopp.univ-montp2.fr/manual/html/bppsuite/2.4.0/bppsuite.html#bppancestor

Cheers,

Laurent

[Liang Wang]

Dear Laurent,

Thank you for your patience. The amino acid substitution model named FLU was presented in the reference (FLU, an amino acid substitution model for influenza proteins). The best fitting model (FLU+I+G) my data was chosed by prottest.  In this best model, "FLU" was the amino acid substitution model, "I" represented extent of static, unchanging sites in our dataset, and "G" indicating gamma distributed rate variation among sites. And I wanted to know how did I set the parameters in config file when I  proceed ML fit of the model/tree on the data with bppml, before I reconstruted ancestor sequence?

Best wishes!

 

[Julien Y. Dutheil]

Hi, 

I'd just like to add that it is also possible to provide an empirical model to bppML/bppAncestor, both programs accept a .dat file in PAML format. If such a file is available for the FLU model, you could use it directly withou having to program anything.

J.

[Liang Wang]

Hi Julien:

     Thank you for your patience. As you mentioned that I can provide a custom empirical model in PAML format for bppML/bppAncestor. Is that mean I can set  model paramter as "model=Empirical(name=FLU,file=path_of_empirical_model/model.dat)" in config file?

     The reference that proposed the FLU substitution model was deposited in ftp://ftp.sanger.ac.uk/pub/1000genomes/lsq/FLU. Is the file named "Flu_All_it2.txt_PAML.txt" a PAML format empirical model which can be directly used in bppML/bppAncestor?

[Laurent Guéguen]

Dear Liang,

it should be fine like this. If you have problems, or if you prefer that the model is integrated in bio++, just tell.

Cheers,

Laurent

[Julien Y. Dutheil]

Hi Liang,

This does indeed look like a model file, and it should be readable by bppSuite, using the option you wrote before. Then you can combine the gamma + invariant model with it, and it should work (otherwise, please let us know)!

Julien.

[Liang Wang]

Dear Laurent,

       

       Thank you for your kind help and patient explanation. I will have a try according to your advice. ^_^

[Liang Wang]

Dear Julien,

 Your suggestion is very helpful to me. I will practice it immediately.