Tide fasta-indexing

[GoogleCodeExporter]

Good Afternoon SearchGUI people

I'm trying to get searchGUI installed/working on one of our HPC clusters so 
that I can run all samples and database searches in parallel, thus reducing the 
time taken. To do this I'm using the commandline interface to run search gui 
with a single search engine with a single mgf file. I'm setting a temp 
directory (using the option -temp_folder ) for each of these runs to irradicate 
any over-writing of files. However, I'm having an issue with Tide. Despite 
setting a temp_folder, it still writes the tide-index files (auxlocs, protix, 
pepix.nomods, pepix.nopeaks) to a folder where the crux binary is located. As 
such when I set several tide searches running they error because one search is 
overwriting the tide-indexes whilst another is trying to read.

It also seems to write ms2 files to the source location of the COMET search 
engine and not to the Tide location.

Is there any way of specifying SearchGUI to direct ALL temp directories/files 
to a specified location, as adding -temp_folder <path-to-directory> does not 
seem to do this?

Many thanks

Craig

Original issue reported on code.google.com by craig.la...@gmail.com on 12 Mar 2015 at 2:25

[GoogleCodeExporter]

Just an after-thought, would it be possible to specify to searchCLI a tide 
option to turn off the tide-index command as well?

Do you think I may encounter similar problems with the other search engines?

Regards

Craig

Original comment by craig.la...@gmail.com on 12 Mar 2015 at 2:34

[GoogleCodeExporter]

The temp folder setting is only for temp folders created and used directly by 
SearchGUI. Most search engines also have their own temp folders, usually either 
in the same folder as the executable or in a sub folder. For some search 
engines it might be possible to change this/these folder/folders, but for 
others this is not possible. You would therefore have to look into the 
documentation for each search engine to figure out which of them supports this.

And if you're only planning on running a single search engine on the command 
line it might be easier to just use the search engine command line directly 
instead of going via SearchGUI? Then you'd also have more control over these 
search engine specific details.

Regarding the location of the ms2 file it is correct that we put this in the 
Comet folder. This is just in order to not have to do the conversion from mgf 
twice. We might look into putting it in the Tide folder if only Tide is used 
though. 

Finally, turning of the tide-index does not make much sense as this is required 
for the tide-search. Or are you referring to then reusing the tide-index and 
therefore don't want to overwrite/recreate it? This would again be very 
specific to the search engine and your setup and I would again argue for rather 
using the search engine command line directly?

The other search engines will have similar issues in your setup I think. We are 
working on a cluster setup of SearchGUI and PeptideShaker which will hopefully 
be made available in the near future.

Original comment by harald.b...@gmail.com on 12 Mar 2015 at 3:42

[GoogleCodeExporter]

Hi Harald

Thanks for getting back to me so soon. Running the search engines individually 
was the other option I was toying with. I was initially resisting writing 
scripts to do all of this when SearchGUI/SearchCLI has it all nicely packaged 
together and didn't see the point of re-inventing the wheel, so to speak. 

I'll go ahead with this now and see how I get on.

Good news on the cluster setups of SearchGUI/PeptideShaker, if you need anyone 
to test them out at any point just let me know, happy to help.

Regards

Craig

ps I don't suppose you guys are working on a quantitation cog to the 
SearchGUI/PeptideShaker pipeline are you?

Original comment by craig.la...@gmail.com on 12 Mar 2015 at 3:56

[GoogleCodeExporter]

I agree that one should usually not attempt to reinvent the wheel, so if you 
figure out how to alter the search engine command lines to only write temporary 
files to a user defined folder we will consider adding this to SearchGUI.

Our cluster setup is currently being tested in-house. Not sure when/how they 
plan to make it available though. But when it does becomes available it will be 
advertised on the PeptideShaker Google Group: 
https://groups.google.com/forum/#!forum/peptide-shaker.

And a new tool for doing iTRAQ/TMT quantification on top of PeptideShaker 
output will be available soon.

Original comment by harald.b...@gmail.com on 12 Mar 2015 at 4:31

[GoogleCodeExporter]

That's great, thanks. Any plans for label free?

Original comment by craig.la...@gmail.com on 12 Mar 2015 at 4:34

[GoogleCodeExporter]

Plans for label free? Maybe. In the near future? No.

That would require working with the raw files (or mzML), as you need access to 
the MS1 spectra. Our current pipeline relies on the simpler and smaller mgf 
files for the spectrum data.

In general moving to label free will require a lot of work.

Original comment by harald.b...@gmail.com on 12 Mar 2015 at 4:44

[GoogleCodeExporter]

Yes, label-free would be a rather large undertaking. Don't mind me though, I 
was just asking for a moon on a stick. It's just annoying that MQ is tied into 
a single search engine that you can't change, despite it being one of the best 
tools for LF.

Original comment by craig.la...@gmail.com on 12 Mar 2015 at 4:49

[GoogleCodeExporter]

Hi Craig,

Sorry to jump in the discussion. Would it be an option for you to install (ie 
copy) the search engines in/alongside the temp folder? You can provide the 
installation folder in your command line. They will then be run from there and 
all indexes will be located there. I guess this would solve your problem?

Best regards,

Marc

Original comment by mvau...@gmail.com on 14 Mar 2015 at 10:52

[GoogleCodeExporter]

Hi Marc

Great minds think alike, this is what I was testing out yesterday. Tested
with 4 so far and works. Not very pretty coding, but it's seems to be
getting the job done.

Original comment by craig.la...@gmail.com on 14 Mar 2015 at 11:44

[GoogleCodeExporter]

(issue closed)

Original comment by harald.b...@gmail.com on 23 Mar 2015 at 6:40

• Changed state: Fixed