Issues in building library

Issues in building library nehasharma ns27  2019-07-11


I a new skyline user. I was able to run the software for the demo files that have been provided with the MS1 tutorial but now when I am trying to work on my data, I am having to issues:

  1. When I am trying to build spectral library from .mzid file, it is showing some errors (PFA: File No.1)
  2. When I am using .DAT format, although library is being built but chromatograms are not being extracted in the next step. There seems some missing files and even after processing the same, chromatograms are not being displayed in the final processing (PFA: File No. & 3)

Please help me with this.


Brendan MacLean responded:  2019-07-11

Hi Neha,
Are you searching with Mascot (as you have DAT format)? How are you producing the .mzid file? Can we get you to upload the .mzid and MGF files to our file drop point to allow us to look at what is going on:

The second screenshot when you are using the DAT file is just telling you that Skyline cannot find the WIFF file you searched to produce the DAT file, and since the DAT file itself does not contain any MS1 spectra, without that file Skyline will not be able to extract chromatograms for you. You can still do that import now after the wizard has completed using File > Import > Results and just pointing Skyline at the WIFF file it was requesting before.

In the future, when you are using the wizard, you should try to make sure you have the original WIFF files available and if you don't want to put them in the same folder as your DAT file or your .sky file then you should be prepared to click either the Find or Find in Folder button shown on the wizard page in your File No. 2 image.

Hope this is enough to get you your first chromatograms from your own data. We would appreciate having a look at your .mzid file. Thanks for posting to the Skyline support board.


nehasharma ns27 responded:  2019-07-11

I have uploaded the .mzid file at your file drop point. Please have a look!!

matt.chambers42 responded:  2019-07-11

Hi Neha,

The problem with the mzIdentML is a small bug in ProteoWizard (what BiblioSpec is using to read the mzid): it expects <affiliations> element to be upper-case <Affiliations> but this expectation is wrong for mzIdentML 1.0. If this was the only problem, then it would be easy to work around by simply opening the file up in a text reader to make the change. The elements are at lines 49 and 52.

However, there is a bigger problem. Not all the <SpectrumIdentificationItem>s have <PeptideEvidence> elements. That's just plain wrong and nonsense. It makes the file pretty much unusable.

In short, you're using a pretty old Mascot Server, it seems they updated the mzIdentML exporter in 2.4, and you will probably have to update to get a fixed mzIdentML exporter. But you'll probably be able to get the DAT reader working with Brendan's feedback above, so it may be a moot point.

nehasharma ns27 responded:  2019-07-12

Thanks Brendan and Matt.

I have successfully built the library. but now there is some issue with the valid retention times. Please have a look the attached file and give your valuable suggestions. It will be a great help.


matt.chambers42 responded:  2019-07-12

Hmm, to troubleshoot this we'll need the DAT file. Shouldn't need the WIFFs just to check why it's not getting retention times.

nehasharma ns27 responded:  2019-07-13

Please find attached the required files at the file drop point of skyline support.


matt.chambers42 responded:  2019-07-16

Did the MGF file you fed to Mascot have RTINSECONDS fields? If not then the retention times won't be available in the DAT file and it's not in the scan titles either. To get retention times from the WIFF files BiblioSpec would need to be modified to try to find the spectra in external files. WIFF is really the worst case for this because its nativeIDs are so complex (e.g. "sample=1 period=1 cycle=4086 experiment=2" corresponds to your scan titles: "Locus:" except the first 1 refers to file number). I'll check with Brendan to check whether we can fix it.

Brendan MacLean responded:  2019-07-16

Here are some notes I wrote about missing retention times, especially in Mascot results:

This was especially prevalent at the time it was written (2013) when using Proteome Discoverer in combination with Mascot. We spent years complaining to various peptide search providers, and at this point, we do not receive a lot of complaints about it.

I think the solution will need to be finding a search workflow that gets the RTINSECONDS into the MGF and eventually the .dat file.

I don't see an attached file, but I suspect, as Matt suggests, that the DAT file lacks the necessary RT information, as described on the page linked above. There should be a way to fix this without requiring a fix in the Skyline library builder (BlibBuild).