LibraPeptideParser takes FOREVER on linux

[ChargedPeptide]

I've recently built TPP 4.6.2 on both Centos 6 and Ubuntu 12.1/13.04.
In all cases I have the same problem, the librapeptideparser step takes literally forever (36 hours and counting) when processing itraq-data.
I've ran this exact workflow on TPP-4.5.2 on a windows box and LibraPeptideParser would take minutes while the bulk of time would be spent processing LibraProteinRatioParser,
though never more than an hour (and that is on a very slow box).
Since LibraPeptideParser gives NO output it is impossible to tell what is going on, are there any logs generated by this?.
If any (devs especially) have any ides, know a global variable that needs to be set or ANYTHING else about this it'd be appreciated.
The mzxml files are converted by msconvert from mgf files output by mascot distiller.
Below is XInteract output for a test run of two files, though probably only the last lin:


running: "/home/erik/tpp/bin/PeptideProphetParser '/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep.xml' MINPROB=0.05 DECOY=dc ACCMASS"
using Accurate Mass Bins
Using Decoy Label "dc".
 (X! Tandem (k-score))
adding Accurate Mass mixture distr
init with X! Tandem (k-score) trypsin
MS Instrument info: Manufacturer: UNKNOWN, Model: UNKNOWN, Ionization: UNKNOWN, Analyzer: UNKNOWN, Detector: UNKNOWN

 PeptideProphet  (TPP v4.6 OCCUPY rev 2, Build 201304281953 (linux)) AKeller@ISB
 read in 42 1+, 6239 2+, 2220 3+, 1566 4+, 0 5+, 0 6+, and 0 7+ spectra.
Initialising statistical models ...
Found 3290 Decoys, and 6777 Non-Decoys
Iterations: .........10.........20.......
model complete after 28 iterations
command completed in 16 sec

running: "/home/erik/tpp/bin/ProphetModels.pl -i /home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep.xml -d dc"
Analyzing /home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep.xml ...
Reading Accurate Mass Model model +1 ...
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Reading Accurate Mass Model model +2 ...
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Reading Accurate Mass Model model +3 ...
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Reading Accurate Mass Model model +4 ...
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Reading Accurate Mass Model model +5 ...
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Reading Accurate Mass Model model +6 ...
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Reading Accurate Mass Model model +7 ...
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Parsing search results "/home/erik/tpp/data/taskController/LibraAllStepsConvert288e7eb159962475699b2a37526bbc8f9/MsConvertplugin/40a1ad168ae341b088edbe0233d45042/output/Batch2SaproItraqFrac100mMa (X! Tandem (k-score))"...
  => Total of 2057 hits.
  => Total of 0 decoy hits.
  => Total of 0 excluded hits.
Parsing search results "/home/erik/tpp/data/taskController/LibraAllStepsConvert288e7eb159962475699b2a37526bbc8f9/MsConvertplugin/40a1ad168ae341b088edbe0233d45042/output/Batch2SaproItraqFrac100mMb (X! Tandem (k-score))"...
  => Total of 3601 hits.
  => Total of 0 decoy hits.
  => Total of 0 excluded hits.
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Warning: empty y range [0:0], adjusting to [0:1]

plot "/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.tsv" using 1:17 title "Observed" with line lc -1 , "/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.tsv" using 1:18 title "Model Pos" with line lc 3 , "/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.tsv" using 1:19 title "Model Neg" with line lc 1
                                                                                                                           ^
"/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.gp", line 23: warning: Skipping data file with no valid points
Warning: empty y range [0:0], adjusting to [0:1]

plot "/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.tsv" using 1:20 title "Observed" with line lc -1 , "/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.tsv" using 1:21 title "Model Pos" with line lc 3 , "/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.tsv" using 1:22 title "Model Neg" with line lc 1
                                                                                                                           ^
"/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep_FVAL.gp", line 25: warning: Skipping data file with no valid points
Warning: empty y range [0:0], adjusting to [0:1]
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 9: reading configurations from ~/.fonts.conf is deprecated.
command completed in 2 sec

running: "/home/erik/tpp/bin/LibraPeptideParser '/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/interact.pep.xml' -c/home/erik/tpp/tmp/6c80b7f63abd436295d2ec4b65277194/condition.xml"

[Luis M]

Hello,

I'll take a look at the code this afternoon, but just for starters: can you verify that your mzXML files have a valid index?  Looking up spectra without the aid of an index will considerably slow down Libra and other tools. 

Cheers,

--Luis

--
You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discu...@googlegroups.com.
To post to this group, send email to spctools...@googlegroups.com.
Visit this group at http://groups.google.com/group/spctools-discuss?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[ChargedPeptide]

Thanks Luis, I really appreciate that.

The problematic files contain this: 

<index name="scan">

    <offset id="0">984</offset>

    <offset id="1">2463</offset>

    <offset id="2">4091</offset>

and a lot of others, so the files seem to be indexed correctly.

Since my last post I've I've tried a windows install of TPP 4.6.2 and have the exact same problem there.

I then built TPP4.5.2 on ubuntu 12.04 and that IS able to run the whole process succesfully but only if starting over all the way back from MSConvert.

So I took the MSConvert version from there, put it with the rest of TPP 4.6.2 and VIOLA, everything works.

So the data is definitley rubbing libraPeptideParser the wrong way but I have no idea how.

I'd be happy to provide anything else you need. 

I'm going to give TPP 4.6.1 a shot with my test-cases and see if I can at least get a semi-modern experience.

Thanks for your work.

/Erik

I'm gonna give 4.6.1 a shot and 

[ChargedPeptide]

Okay, I realized that MSConvert from TPP 4.5.2 just ignored my  sortByScanTime filter for msconvert, tried this with 4.6.2 and it works. So it seems I might have to give up my sorted scans. Sniff.

[Luis Mendoza]

Glad that you found the source of the issue and thanks for reporting back.  Not sure why Libra would be slow, other than the index being wrong after applying the filter.  (e.g. can you notice a lag when opening a spectrum in the spectrum viewer?)  Perhaps something to take up with the msconvert/ProteoWizard team.

Cheers,

--Luis

On Mon, Apr 29, 2013 at 3:46 PM, ChargedPeptide <malm...@gmail.com> wrote:

Okay, I realized that MSConvert from TPP 4.5.2 just ignored my  sortByScanTime filter for msconvert, tried this with 4.6.2 and it works. So it seems I might have to give up my sorted scans. Sniff.

--

[Joseph Slagel]

Just an FYI, I took a quick look at the difference between msconvert in TPP 4.5.2 and 4.6.2 and it looks as if the version in 4.5.2 doesn't support the sortByScanTime filter.  I'm going to assume that this is a relatively new filter added by the ProteoWizard folks.

Any chance you can provide the mzML files created using msconvert from 4.5.2 and 4.6.2 so that we can take a look to see what may be wrong?

-Joe

On Mon, Apr 29, 2013 at 3:46 PM, ChargedPeptide <malm...@gmail.com> wrote:

Okay, I realized that MSConvert from TPP 4.5.2 just ignored my  sortByScanTime filter for msconvert, tried this with 4.6.2 and it works. So it seems I might have to give up my sorted scans. Sniff.

--

[ChargedPeptide]

Certainly, though I assume the same file processed with/without sortByScanTime would be the best difference to look at. How do I get them to you?

They are small enough to just put in an e-mail if needed.