Thank you for creating HyPhy.
I am trying directionalREL.bf with Win7 x64 and WinXP x32 platforms and
faced error messages as follows,

after > Reload/New (New analysis) > (file specification) > Substitution Model (any model in the list).
------------
"Problem occurred in line:global freqParam_4=vf[k]
Call stack
4 : global freqPram_4=vf[k]
3 : ExcuteCommands in string "global freqParam_"+k+"=vf[k]; freqParam_"+k+":<1;" reading input from
x:\HYPHY2.0\TemplateBatchFiles\TemplateModels\
2 : ExecuteAFile fro Current Task has been terminated. Would you like to see the remaining error messages, if there are any?"

"Problem occurred in line:global freqParam_4=vf[k]
Call stack
4 : global freqPram_4=vf[k]
3 : ExcuteCommands in string "global freqParam_"+k+"=vf[k]; freqParam_"+k+":<1;" reading input from
x:\HYPHY2.0\TemplateBatchFiles\TemplateModels\
2 : ExecuteAFile f Current Task has been terminated. Would you like to see the remaining error messages, if there are any?"

"Problem occurred in line:promptModel(0)
Call stack
1 : promptModel(0)
Current Task has been terminated. Would you like to see the remaining error messages, if there are any?
------------
Could you kindly confirm the same problem is reproduced in your Windows platform? :'(

---

Dear Kohei,

Thanks for reporting this error message. I confirmed that it is generated if you provide a nucleotide alignment as input to DirectionalREL.bf
The analysis expects a protein (i.e. 20 residues, not 4) alignment, but does not report the error in an informative way. I will fix the file to specifically check for whether or not a protein alignment has been submitted. In the meantime, please try to run DirectionalREL.bf on a protein alignment and tell me if it works on your system.

Sergei

Dear Sergei,

Thank you for the prompt reply!
Now I succeeded to start the calculation (now it's running).
I will check the result when it's finished.

Kohei

Sorry again, but after 37 hours calculation with Core 2 Quad Q9650 CPU (3.00 GHz) machine, another error message appeared.

---

This is another message,

---

and this is HYPHY Console Window.

---

and this is the data file analyzed. Did I have to remove questionable sites (?) and sequences with too many blanks?

---

This looks like an execution path problem.  Where is your HYPHY distro on your hard drive?  Where are the DREL files?  

Thank you for the response.
As in the error messages, the HYPHY distributions are under K:\MyDocuments\Downloads\HYPHY2.0

. What is "DREL files"? If it is the data file being analyzed, K:\MyDocuments\Documents\n中込研\YenProject\20100408~HyPhy解析\20100625DirectionalEvolutionProteinSequences(DEPS)を試す

oh...problem in the folder name?

Based on the console output, it looks like the path is okay.  Can you check the folder HYPHY_BASE_DIRECTORY/ChartAddins/DistributionAddIns/Includes to make sure that the posteriors.ibf include file is present?

Oh, and DREL is another name for DEPS.  
Ambiguous and partial sequences should be OK.
I'm wondering if this is a locale issue (i.e., directory separators).  Sergei?

I found posteriors.ibf under K:\MyDocuments\Downloads\HyPhy1.0\HYPHY_Win32\ChartAddIns\DistributionAddIns\Includes

but the issue is I found no posteriors.ibf under K:\MyDocuments\Downloads\HYPHY2.0
, which I am using. Is this the reason?

Hi Kohei,

The 2.0 windows distribution was missing the ChartAddIns directory. I have just rebuild it: please download it from Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login and see if this fixes the missing file issue. Try running a small protein file first to confirm that everything is in order.

Sergei

Thank you for the quick fix. I am now re-running the analysis and will inform you the results.

Thank you for your help.
First, DirectionalREL.bf was run with JTT model for the same alignment, and the calculation finished within 5 hours. Amino acid sites 51, 120, 157, and 309 were selected (please see attached file). However, I found so many N/A in a resulted xxx_AA_bysite.csv file and I doubted whether the settings were adequate. Are there any documentation about DirectionalREL.bf in addition to Kosakovsky Pond et al. (2008) Molecular Biology and Evolution 25:1809-1824?

---

Second, I re-ran DirectionalREL.bf with REV model; it took about 33 hours. Now amino acid sites 51, 125, 236, and 242 are selected; these sites are "more preferable" in that they correspond with variable regions known in the gene, which encodes viral outer capsid. However, I still find so many N/A in the resulted xxx_AA_bysite.csv file. Does this situation occur when the variation (the number of substitutions) in the alignment is too small ?

---

Hi Kohei,

N/As are normal. As described in the DEPS manuscript, the stage consists of two steps:

1). For each target residue, a p-value is computed to determine if there are ANY sites (proportion > 0) directionally evolving towards that residue. That's the top row of the spreadsheet that you attached.

2). For those residues which show significant p-values (after a multiple test correction), Bayes Factors are computed

N/A simply means that there was no need to compute Bayes Factors for that particular residue/site pair because the LRT test was not significant.

Sergei

Dear Sergei,

Thank you for your kind reply.
I should have read the DEPS manuscript in detail; now I am trying to do so. I hope that I can obtain meaningful information.

Kohei