[quote author=konrad  link=1129736474/0#0 date=1129736474]

1. What is the easiest way of incorporating one of the models in PositiveSelectionLI.bf into a script such as presented in the tutorials? Also, what is the easiest way of inferring and using an ML tree in such a script?

--------

2. On running PositiveSelectionLI.bf (the help documentation refers only to PositiveSelection.bf, which I do not find. One difference seems to be that the documented "single rate" option is not available - are the help files outdated?) I get an error after specifying the rate distribution:

Error:
Bad symbols in expression rate4Leaves'?
Current BL Command:Build Formula:rate4inodes=rate4Leaves'

Check errors.log for details on execution errors.

Check messages.log details of this run.
Segmentation fault

-------

[/quote]

Standard analyses documentation is [b]very[/b] out-of-date (sadly). PositiveSelection.bf has been renamed to NielsenYang.bf and PositiveSelectioLI.bf is an experimental (unsuccessful) analysis to test for differential selection between leaves and internal nodes of a tree.

In terms of incorporating standard models.

You can use SelectModel (datafilter); to interactively select models defined in TemplateModels/ (models.lst holds the list of models presented to the user), or use the #include command to read in outboard file definitions.

NielsenYang.bf has a PopulateModelMatrix function to set the cells of a rate matrix (controlled by the modelType flag) and SetWDistribution () to set the distribution of \omega to one of the pre-defined ones. You might be better off copying NielsenYang.bf and modifying it directly if you want to use the same set of models.

[quote]

3. On running SingleBreakpointRecomb.bf or SingleBreakpointRecomb-2.bf (what is the difference between these two?) I get an error after specifying the model I want to use:

Please type in the abbreviation for the model you want to use:GY94
Error:
Could not read batch file:TemplateBatchFiles/TemplateModels/TemplateBatchFiles/TemplateModels/GY94.md
l
Current BL Command:Select Template Model for {filteredData}

Check errors.log for details on execution errors.

Check messages.log details of this run.
Segmentation fault
[/quote]

'-2' again is an experimental analysis.

This seems like a BASEPATH issue again, try to rerun with BASEPATH=/pathtohyphy/ on the command line and see if this helps.

Cheers,
Sergei

Dear Konrad,
[quote author=konrad  link=1129736474/0#2 date=1130246542]

The modelType flag is set by chooseGeneticCode.def, where it is used to identify the genetic code. But in PopulateModelMatrix it seems to mean something else: kappa is only used for modelType = 2 and 3.

[/quote]

Before 'PopulateModelMatrix' is called, 'modelType' (I don't think PopulateModelMatrix uses 'modelFlag') is set by the user response to a dialog:

[code]ChoiceList (modelType,"Choose a model",1,SKIP_NONE,
                 "MG94 1x4","Muse-Gaut 94 model with 4(-1) nucleotide frequency parameters (intra-codon position independent).",
                 "MG94 3x4","Muse-Gaut 94 model with 12(-3) nucleotide frequency parameters (intra-codon position specific).",
                 "GY94 1x4","Goldman-Yang 94 model with 4(-1) nucleotide frequency parameters (intra-codon position independent).",
                 "GY94 3x4","Goldman-Yang 94 model with 12(-3) nucleotide frequency parameters (intra-codon position specific).",
                 "MG94 Custom 1x4", "Muse-Gaut 94 crossed with an arbitrary nucelotide substitution model with 4(-1) nucleotide frequency parameters (intra-codon position independent).",
                 "MG94 Custom 3x4", "Muse-Gaut 94 crossed with an arbitrary nucelotide substitution model with 12(-3) nucleotide frequency parameters (intra-codon position specific)."
); [/code]

Since selection is zero based, you will see that agreement with PAML (which uses GY94 for most analysis) is only to be expected for [tt]modelType == 2[/tt] or [tt]modelType == 3[/tt].

Sorry for my undocumented code. Hopefully this helps.

Cheers,
Sergei