[BioPython] Blastall problem w/ restrict_gi

[Peter]

Roger Barrette wrote:
> Hello, I'm new to the list, and relatively new at Python.

Hi Roger, and welcome to the list!

> I need to run local blast using tblastx, but I have to limit my 
> searches to subsets of my local database.  To do this I have gi lists
>  (*.gid.txt file) obtained from NCBI, to define my subsets.  To run 
> this blast, I'm using the following command to run blastall in my 
> script:
> 
> result_handle, error_info =
> NCBIStandalone.blastall("/BLAST/blastall.exe", "tblastx", 
> "/BLAST/DATAout/VirDBX", "/BLAST/sequencesXX.fasta", "7",
> restrict_gi="/BLAST/DATAout/10241.gid.txt")
> 
> When I include the restrict_gi keyword and option, I get no results 
> back when I run this through python.

Could you be a little more specific about what goes wrong? Also are you
using Windows, what version of Biopython and what version of Python?

Have you looked at the contents of both result_handle AND error_info?
You say you get no results back (is result_handle is blank?), so
checking error_info would be a good idea.  Try something like this...

save_file = open("my_blast.xml", "w")
save_file.write(result_handle.read())
save_file.close()

save_file = open("my_blast.err", "w")
save_file.write(error_info.read())
save_file.close()

> I went into NCBIStandalone and modified it to print out the command 
> that is supposed to be passed through the os.popen3() command, which
>  is:
> 
> /BLAST/blastall.exe -p tblastx -d /BLAST/DATAout/VirDBX -i 
> /BLAST/sequencesXX.fasta -m 7 -l /BLAST/DATAout/10241.gid.txt
> 
> When I copy this string directly into the windows command line, I get
> results, and it works fine, but it doesn't work when called through
> python. It does work in Python , however, if I don't include the 
> "restrict_gi" option.    Can anyone suggest a modification to the 
> Blastall function or how I call blast from my script that may fix 
> this problem?

Have you tried running this command at the command line, and redirecting
the output to a file (e.g. test.xml) and then getting Biopython to parse
that file?

i.e. This should tell us if there is a problem parsing the XML output,
or a problem in calling standalone blast.

Peter