Proteome discoverer vs Skyline results

support
Proteome discoverer vs Skyline results Ehklyq23  2025-02-10 13:18
 

I have noticed a pattern where when doing a PRM analysis for a peptide in a complex sample, the peptide is not identified by proteome discoverer, but the peptide transition ions are found in skyline with an idotp > 0.9. Is this a reliable identification of the peptide? It is also observed at the correct retention time based on PRM of a positive control sample.

Is there a reason it is found in skyline but there is not a positive hit in proteome discoverer?

 
 
Nick Shulman responded:  2025-02-10 14:00
In general, a good isotope dot product match is not a very reliable identification of the peptide. Peptides all have about the same proportion of Hydrogen, Oxygen, Carbon and Nitrogen so they have approximately the same isotope distribution. Getting a good idotp score is equivalent to being confident that you have the correct mass and charge state.

It would be much better if you could see that you have a large number of co-eluting MS2 chromatograms because that would tell you that the thing you are observing has the same fragment ions as your peptide of interest.

If you would like, you could send us your Skyline document and we can take a look.
In Skyline you can use the menu item:
File > Share
to create a .zip file containing your Skyline document and supporting files including extracted chromatograms.
Files which are less than 50MB can be attached to these support requests. You can always upload larger files here:
https://skyline.ms/files.url

-- Nick
 
Ehklyq23 responded:  2025-02-10 14:19
Thank you Nick! Please find the attached file.

For an example, I would look at file number ECAZ_QE021789 (the last one). The peptide VGNILQLGSK is not identified in proteome discoverer, but there is a prominent peak in skyline. By prominent, I mean the transition ions are identified at the correct location as well. In contrast, the peptide in positive control it was identified in PD (ECAZ_QE021784).

The file was large and submitted through the link.
 
Nick Shulman responded:  2025-02-10 15:24
Yes, that is the correct peak.
It is about 30 times less abundant than in your positive control.

If you look at the Library Explorer in Skyline you can see a summary of how many peptides were found by your peptide search engine in each of the files that were searched.
You can get to it with "View > Spectral Libraries" and then in the Spectral Library Explorer you can push the "..." button next to the library name.
The Library Details window shows that zero peptides were identified in ECAZ_QE021789.raw.
It is unusual to see such a zero like that. I wonder whether something went wrong in your peptide search and Proteome Discoverer was not able even to open the file "ECAZ_QE021789.raw" and that was the reason that it did not detect any peptides.

By the way, I see that you are using Skyline 23.1.
The latest released version of Skyline is 24.1. We always recommend using the latest version of Skyline unless you are trying to reproduce results from old experiments.
Nothing much has changed in this area in recent versions of Skyline, but the newer versions are always better.
-- Nick
 
Ehklyq23 responded:  2025-02-10 15:57
Hi Nick,

Thank your for your speedy response. I'm sending the full analysis which might make things more clear. I edited the file but that might have resulted in some missing data. In this file it is showing that ECAZ_QE021789.raw has peptides identified.

This also gives a bit more context. In each raw file a good peak for VGNILQLGSK is observed at the correct retention time. However, it is only identified in proteome discoverer in less than half of them. My question is which is more believable, the proteome discover results or the observed peak in skyline? Looking at the 'Peptide IDs' they're only observed in the PD positive files. Is that the cause of the difference?

Thank you for your time and patience answering my questions!
 
Nick Shulman responded:  2025-02-11 16:37
Peptide search results should not be used to infer the absence of a peptide.

That is, you might say "Proteome Discoverer did not find the peptide in this sample" but that is not the same as "Proteome Discoverer said that this peptide is not present in this sample".
It is likely that Proteome Discoverer found rather good evidence of the peptide in that sample, but the evidence was not strong enough to be reported with a 1% false discovery rate cutoff.

Once a peptide has been detected in one of your samples, you can assume that the peptide is present in all samples, although it might be present at an amount that is below the limit of detection.
If you integrate the chromatogram peak area at the place where you expect the peptide to be found, you can get an upper bound on the quantity of peptide that might be present.

If you are interested in comparing quantities between groups of samples, and determining which peptides are present in different amounts in the two groups, we recommend the Group Comparison tutorial:
https://skyline.ms/wiki/home/software/Skyline/page.view?name=tutorial_grouped
-- Nick