For those of us who are either intimidated by the exigencies of designing substitution models in HyPhy or just feel incompetent, is there any chance that the number of available models could be increased to parallel those available in, let's say ModelTest?

Thanks.

Dear Alan,

HyPhy actually implements ModelTest.
Look under Standard Analyses->Model Comparison->ModelTest.bf
Also, Standard Analyses->Model Comparison->NucModelCompare.bf lets you test all models nested within the GTR model. Any other nucleotide standard analysis has the CUSTOM model option which lets you specify any of the 203 (and optionally +G +I and many other types of rate variation) nucleotide models.

If you are working through the data panel interface, you can always select GTR and then constrain it down (e.g. to make K3ST etc). Refer to Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login for instructions on how to constrain parameters.

Or are you looking to simply have more predefined models in the drop down menu in the data panel interface?

Cheers,
Sergei

Dear Sergei,

Yes, I am "looking to simply have more predefined models in the drop down menu in the data panel interface."

Best,
Alan

Dear Alan,

Right. I wish I had more time... :(

Can you list the models you most would like to see? ModelTest has 8 rate matrices (JC69, HKY85, TrN93 etc), the other model options are base frequencies and site to site rate variation -- you can stack those on top of the basic rate matrix to generate HKY85+G+I for example.

Cheers,
Sergei

Dear Sergei,

I understand.  I feel a bit churlish asking authors of useful, free software to do more work at my behest!  As a newcomer to HyPhy, I am not yet in the comfort zone in programming alternative subst. models.

SYM, TIM, and TVM would be nice, but I can also accept that I need to gird my loins and develop the parameter files for them myself.  In the custom mode, can one just input the numerical substitution code?  If it is as easy as that (plus specifying rates and base frequencies), I guess I shouldn't have even asked the question.

Best,
Alan

Hi Alan,

Try running AnalyzeNucProtData.bf (Standard Analyses->Basic Analyses) on an alignment. When prompted for the model, select 'CUSTOM' then enter the six character model code, select Global if you want no rate variation and Global w/variation for further options (like G or G+I), rate parameters (estimated or fixed) and base frequencies (observed, equal or custom).

That should cover all the bases:) The CUSTOM model should be available for all nucleotide data analyses accessible via the Standard Analysis route.

Cheers,
Sergei