Trouble importing peptide search ("Importing the FASTA did not create any target proteins")

support
Trouble importing peptide search ("Importing the FASTA did not create any target proteins") whitsonj  2017-09-25
 
Hi, I'm a new Skyline user and I'm having trouble getting some global phospho proteomics data into Skyline.

I ran our .RAW phospho data through ProteoWizard and the Crux-pipeline in order to produce the peptide search files in .SQT format as well as a .perc.xml file. I've been trying to import the search into Skyline using the import peptide search wizard, following the instructions in the MS1 filtering tutorial. When I get to the last step Skyline tells me that "Importing the FASTA did not create any target proteins" and will not proceed with the import. I'm just using the most up to date Uniprot FASTA of all mouse proteins. I also have the associated .MS1, .MS2, and .RAW files in the same directory with the .SQT and .perc.xml files and the names all match up.

Any help would be greatly appreciated!
 
 
Kaipo Tamura responded:  2017-09-25
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
 
whitsonj responded:  2017-09-25
Kaipo,

I have finished uploading the files. Thanks.

-Jeremy
 
Kaipo Tamura responded:  2017-09-25
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
 
whitsonj responded:  2017-09-25
Kaipo,

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

-Jeremy
 combined-results.perc.xml   16694_mmp_mouse_SS31_H_IMAC_Y5.sqt   16695_mmp_mouse_SS31_H_IMAC_1042-5.sqt   16697_mmp_mouse_SS31_H_IMAC_Y6.sqt   16707_mmp_mouse_SS31_H_IMAC_1039-5.sqt   16708_mmp_mouse_SS31_H_IMAC_1039-1.sqt   16709_mmp_mouse_SS31_H_IMAC_1040-2.sqt   16710_mmp_mouse_SS31_H_IMAC_1042-1.sqt   16711_mmp_mouse_SS31_H_IMAC_1041-1.sqt   16712_mmp_mouse_SS31_H_IMAC_1040-5.sqt   16713_mmp_mouse_SS31_H_IMAC_1042-4.sqt   16714_mmp_mouse_SS31_H_IMAC_Y1.sqt   16715_mmp_mouse_SS31_H_IMAC_1040-1.sqt   16716_mmp_mouse_SS31_H_IMAC_Y2.sqt   16717_mmp_mouse_SS31_H_IMAC_1040-3.sqt   16719_mmp_mouse_SS31_H_IMAC_Y3.sqt   16720_mmp_mouse_SS31_H_IMAC_1040-4.sqt   16721_mmp_mouse_SS31_H_IMAC_1039-3.sqt   16722_mmp_mouse_SS31_H_IMAC_1041-3.sqt   16723_mmp_mouse_SS31_H_IMAC_Y4.sqt   16724_mmp_mouse_SS31_H_IMAC_1039-2.sqt 
 
Kaipo Tamura responded:  2017-09-25
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
 
Mike MacCoss responded:  2017-09-25
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
 
whitsonj responded:  2017-09-25
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
 
Brendan MacLean responded:  2017-09-25
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.
 
Mike MacCoss responded:  2017-09-25
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