Would you mind uploading the .blib file that the Import Peptide Search Wizard creates, along with the FASTA file to: https://skyline.ms/files.url

Thanks,
Kaipo

Kaipo,

I have finished uploading the files. Thanks.

-Jeremy

Hi Jeremy,

It looks like something may have gone wrong when building the spectral library - it only contains spectra for four peptides:
GALDLM[+16.0]LQVNMTPGHSSAPPKETSIGILSAAVSRLEQTPMPNM[+16.0]FGGG
GNPELSPNNLM[+16.0]RPLLNYGIAC[+57.0]MSMGFLVEETAPLVWRGLMVM[+16.0]S
LPGKPSADESQAVQIHM[+16.0]GLALGM[+16.0]FLSRLC[+57.0]EEKLSDM[+16.0]SGQQM[+16.0]
VGAAAAVAFGSGALMLGM[+16.0]FVLQLGVLSTFLSEPVIKALTS

Looking at the Percolator results file may provide more information - maybe all of the other PSMs had q-values that did not meet the cutoff threshold? If you'd like, I can have a look at the sqt/perc.xml (you can attach to this message, or upload them at the previous URL).

Thanks,
Kaipo

Kaipo,

Thanks for looking into this. I've attached the SQT and perc.xml files that the crux-pipeline generated.

-Jeremy

Hi Jeremy,

It looks like though there were many PSMs, Percolator only assigned good q-values to four of them. Can you try rebuilding the spectral library with a cutoff score of 0, then looking at the spectra in the library explorer (View -> Spectral Libraries)? If the spectra look reasonable (many transitions annotated), then something may be going wrong with Comet or Percolator.

Thanks,
Kaipo

Jeremy,
Where did these data come from? The files are labeled IMAC but you are not including phosphorylation in your database search. Just looking at the header of the SQT file shows that you are only searching for oxidized methionine.

H    SQTGenerator    Comet
H    Comment    CometVersion 2015.01 rev. 2
H
H    StartTime    09/21/2017, 03:44:07 PM
H    EndTime    09/21/2017, 03:44:07 PM
H
H    DBSeqLength    0
H    DBLocusCount    0
H    FragmentMassMode    AMU
H    PrecursorMassMode    PPM
H    SQTGeneratorVersion    N/A
H    Database    /net/pr/vol1/ProteomicsResource/dbase/UniProt/20160308/MOUSE.fasta.20160308
H    FragmentMasses    MONO
H    PrecursorMasses    MONO
H    PrecursorMassTolerance    10.000000
H    FragmentMassTolerance    0.020000
H    EnzymeName    Trypsin
H    Algo-IonSeries    0 0 0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0
H    DiffMod    M*=+15.994915
H    StaticMod    C=160.030649

It would be good to get a bit of a sense of what you were trying to do and what the data are. Also what sort of instrument are these data from? If these are Orbitrap-Orbitrap data then I'm not sure why the FragmentMass are used as AMU as opposed to PPM.

Anyway, it would probably be a good idea to revisit the search parameters.

-Mike

Kaipo,

Thanks for the suggestion. Setting the cutoff to 0 did allow me to complete the import and the spectral library explorer shows almost 8,000 peptides.

Mike,

These data were generated by Judit Villen's lab using their Orbitrap Velos instrument. We are trying to determine differentially phosphorylated proteins among several different groups of mice. We are trying to QC the data using Skyline, as that has been helpful to us when working with other phospho proteomic datasets in the past.

I didn't realize that the peptide search was looking for oxidized methionine and not phosphorylation. Could you advise on how to change those parameters when submitting the search? This is my first time using the crux-pipeline so I apologize for my ignorance. Thanks for the help.

-Jeremy

Note that we do not recommend using a 0 cutoff in this case. I hope it was useful for diagnosis, but we only recommend using 0 as a cutoff when the results have already been filtered to some known false discovery rate by another tool. Some tools allow you to do this and then export the results to a format like pepXML, which Skyline will accept.

Otherwise, if you use zero as a cutoff, you are likely accepting FDR of 50% or higher.

Sounds like Mike has helped you onto the right track for a valid Comet search where Percolator will likely find more results below a reasonable FDR cutoff.

Hi Jeremy,
I would discuss this with the Villen lab. There is more to identification of phosphopeptides ... you also need to compute a localization score. Currently there isn't an equivalent to Ascore in Crux but the Villen lab does this on a regular basis. I would discuss this with them to do the search. From there we can help you with the library building step.

-Mike