• Sign In
  1. Messages

Sequest cluster problem

CPAS Forum (Inactive)

View Message

view list print
Sequest cluster problem eunokpaek  2007-04-11 01:08
Status: Closed
Assigned To: wnels2
 
We are trying to use SEQUEST Cluster on CPAS.
I installed labkey server using binary installer.
Also I downloaded and installed SequestQueue on the master node of my Sequest Cluster.
I tested Sequest search on CPAS for a mzXML file, and got such error message in log file.
As you know, the master node of Sequest Cluster is based on Windows, and other nodes are based on Linux.
So, I think this error is from difference path for database file between windows and linux.
I checked my sequest.params and Sequest_default_input.xml file but could not find this path in these files.
Please let me know how to solve this problem.

Thanks.

11 4월 2007 16:15:47,562 INFO : Sequest session ended.
11 4월 2007 16:15:47,562 INFO : Converting Sequest summary file to pepXML.
11 4월 2007 16:15:47,578 INFO : running: Sequest2Xml Enolase_020206_II_15.html
11 4월 2007 16:15:47,796 INFO : params file here: sequest.params
11 4월 2007 16:15:47,843 INFO : creating tgz archive of dta.
11 4월 2007 16:15:47,843 INFO : running: bsdtar.exe czf C:\Documents and Settings\MASS\바탕 화면\신병희_200403_200702\data\Chlorinated Std Peptides\test_search\sequest\test2\Enolase_020206_II_15.work\Enolase_020206_II_15.pep.tgz *
11 4월 2007 16:15:48,968 INFO : xinteract output
11 4월 2007 16:15:48,968 INFO : =======================================
11 4월 2007 16:15:48,968 INFO : running: xinteract -Opt -NEnolase_020206_II_15.pep.xml C:\Documents and Settings\MASS\바탕 화면\신병희_200403_200702\data\Chlorinated Std Peptides\test_search\sequest\test2\Enolase_020206_II_15.work\Enolase_020206_II_15.xml
11 4월 2007 16:15:49,109 INFO :
11 4월 2007 16:15:49,109 INFO : xinteract (TPP v2.9 GALE rev.9, Build 200703101511(Win32))
11 4월 2007 16:15:49,109 INFO :
11 4월 2007 16:15:49,109 INFO : running: "InteractParser "Enolase_020206_II_15.pep.xml" "C:\Documents and Settings\MASS\바탕 화면\신병희_200403_200702\data\Chlorinated Std Peptides\test_search\sequest\test2\Enolase_020206_II_15.work\Enolase_020206_II_15.xml""
11 4월 2007 16:15:49,390 INFO : file 1: C:\Documents and Settings\MASS\바탕 화면\신병희_200403_200702\data\Chlorinated Std Peptides\test_search\sequest\test2\Enolase_020206_II_15.work\Enolase_020206_II_15.xml
11 4월 2007 16:15:49,500 INFO : processed altogether 201 results
11 4월 2007 16:15:49,750 INFO : command completed in 0 sec
11 4월 2007 16:15:49,765 INFO :
11 4월 2007 16:15:49,765 INFO : running: "PeptideProphetParser "Enolase_020206_II_15.pep.xml""
11 4월 2007 16:15:49,921 INFO : (SEQUEST)
11 4월 2007 16:15:49,921 INFO : init with SEQUEST trypsin
11 4월 2007 16:15:49,984 INFO : MS Instrument info: Manufacturer: ThermoFinnigan, Model: LTQ FT, Ionization: ESI, Analyzer: Ion Trap, Detector: EMT
11 4월 2007 16:15:49,984 INFO :
11 4월 2007 16:15:49,984 INFO : PeptideProphet v3.0 April 1, 2004 (TPP v2.9 GALE rev.9, Build 200703101511(Win32)) AKeller@ISB
11 4월 2007 16:15:49,984 INFO : read in 99 1+, 102 2+, and 0 3+ spectra
11 4월 2007 16:15:49,984 INFO : iteration 1 2 3 model complete after 4 iterations
11 4월 2007 16:15:50,125 INFO : command completed in 1 sec
11 4월 2007 16:15:50,125 INFO :
11 4월 2007 16:15:50,140 INFO : running: "DatabaseParser "Enolase_020206_II_15.pep.xml""
11 4월 2007 16:15:50,203 INFO : command completed in 0 sec
11 4월 2007 16:15:50,203 INFO :
11 4월 2007 16:15:50,203 INFO : running: "RefreshParser "Enolase_020206_II_15.pep.xml" "/usr/database/ipi.HUMAN.v3.01.fasta""
11 4월 2007 16:15:50,328 INFO : error: could not open \usr\database\ipi.HUMAN.v3.01.fasta file!
11 4월 2007 16:15:50,343 INFO : - Building Commentz-Walter keyword tree... - Searching the tree...
11 4월 2007 16:15:50,343 INFO : command "RefreshParser "Enolase_020206_II_15.pep.xml" "/usr/database/ipi.HUMAN.v3.01.fasta"" exited with non-zero exit code: -1
11 4월 2007 16:15:50,343 INFO : QUIT - the job is incomplete
11 4월 2007 16:15:50,343 INFO :
11 4월 2007 16:15:50,343 INFO : command "RefreshParser "Enolase_020206_II_15.pep.xml" "\usr\database\ipi.HUMAN.v3.01.fasta"" failed: Unknown error
11 4월 2007 16:15:50,343 ERROR: Failed running xinteract.
 
 
wnels2 responded:  2007-04-11 05:49
Thanks for trying the SequestQueue. On the development system all of the nodes were windows, so please be patient and I'm sure we can get it going.
It looks like most of the stuff is working except that the returned summary.html file still has the Cluster database path in it.
The search is started by sending the remote.params file and the mzXML file to the SequestQueue. There are also parameters sent in the request, one of which is the local CPAS database path. The remote.params file just has the database name, no path. When the remote.params file gets to SequestQueue the path is added; it uses the databaseDir path that is in SequestQueue's web.xml. After the search is completed it converts the sequest .out files to a summary.html file. It then should switch the path of the database in the summary.html back to the CPAS path, using the path that was sent in the request, before it sends it back to CPAS. I think this last step is failing.

I think it would more efficient if I could talk with you on the phone. If you send me an email I'll send you a contact number.

Thanks,
Bill
wnels2@uky.edu
 
wnels2 responded:  2007-04-11 17:05
I noticed that your message came in at 4:00 AM so, in case our time difference makes it difficult, here are some more things you might try. If you go directly to http://SequestClusterServerName/SequestQueue there are some test pages.
1. Select the link for Upload and submit a file for search.
2. For the sequest.params field browse to the pipelineRoot/folder/sequest/protocolName/remote.params file for the sample that failed. Open this file in an editor. It should have just the database name, with no path.
3. For the mzXML file field browse to the pipelineRoot/folder/mzXML file.
4. Click upload file.
5. On the Sequest server in the webapps/SequestQueue directory there are two directories, input and output. Open the sequest.params file in the input directory. The database should be the database from the selected remote.params file plus the database path of the Sequest Cluster.
6. In the output folder open the test1.txt file. Search for the database name. It should be the same as the database in the remote.params (There is no client database path sent in the test, so it shold be the same as it started).

I looked at the Sequest Cluster install manual for the linux nodes. I don't understand why the /usr/database/ is in the returned summary file. I am expecting D:\database

Also, here should have been a warning message, but just in case, the databases on the Sequest Cluster need to be in the CPAS project's database folder (the prophets need them). If you are using indexed databases, a file with the same name as the .hdr file needs to be in the CPAS project's database folder.

Let me know your obsevations.

Thanks,
Bill
 
wnels2 responded:  2007-04-17 05:39
Thanks for testing the patch. I'm glad it worked. I'll post the patch to the Labkey download site.