Major neutral loss peaks not extracted

Major neutral loss peaks not extracted michael plank  2019-01-20

Hi Skyline-team,
can you please help me with this: I performed a PRM experiment for some phosphopeptides, searched them in MaxQuant and generated a library from the MSMS.txt. I loaded phosphopeptides of interest from the library explorer into the target list. These contain positional isomers. Then I imported the data.
When I select e.g. peptide ‘K.INRTRTMSVFDNVSPFKK.T [52, 69] (missed 1)’ the library spectrum is dominated by b-ions with phosphate loss (see sc2.png attached), but the EICs are of the much less abundant y-ions (sc4.png). In ‘Transition Settings – Filter – Ion types’ ‘b’ is included (sc3.png). Is there anything I need to specify that the neutral loss is considered?
(This is also the case when I look a library spectrum of the same sample as the EICs and the RTs match, so its not simply because Im looking at the wrong peak.)
The peak areas view shows the discrepancy between library spectrum and integrated peaks (sc5.png).
The MS2 spectra in QualBrowser reflect the relative ion-intensities as in the library spectrum, apart that the library spectrum is charge deconvoluted.
On an unrelated note: When I change transition settings and then try to re-import data via ‘manage results’ I get the error in sc1.png even though the data have all been imported at this point. Am I doing anything wrong?

Thanks for your help.

Brendan MacLean responded:  2019-01-20

Hi Michael,
Can you show a bit of the Targets list and your Peptide Settings - Modifications? If you have the default Phospho (ST) modification, this should come with the -80 modification defaulting to matching your library spectra. On the other hand, your fragment ion list looks so long, perhaps you have chosen a way of filtering fragment ions that does not rely on your library. Please show your Transition Settings - Library tab also. If you really, really just want to include everything for extraction, then you need to edit the -80 loss in the Phospho (ST) modification to apply even when not matching a library. Because this can be combinatorically quick expensive and result in quite a lot of targeted fragments the default is only to include loss transitions when matching with a spectral library. This is controlled by "Include loss by default" setting, when you select the loss and click the edit button with the pencil icon in the modification editing form. It is set to "Matching Library" by default, but it can also be set to "Always".

Hope this clarifies somewhat, but if you provide the screenshots I have requested, I may be able to offer more.

Thanks for using Skyline in your phospho experiment.


michael plank responded:  2019-01-21

Hi Brendan,

thanks a lot for the prompt help. I really appreciate it.

Please see attached the requested screenshots.
The data above were indeed from a method where I had unticked the 'If a library spectrum is available, pick its most intense ions', but I think the reason for doing that was that I had gotten this unexpected result also with these settings, so I tried disabling the library match.
I now went back to including the library fragment ions and re-imported the results in 'Manage Results', but the problem persists (this is the state where I took the screenshots).
(Also I`m surprised that skyline tries to extract 9 fragment traces while I specified 6 product ions from filtered ion charges and types plus filtered product ions.)
In case there is no obvious reason for this behaviour - maybe it has to do with the problem MaxQuant is currently having with phosphates which we discussed in a separate request?

Thanks again,

Brendan MacLean responded:  2019-01-21

Hi Michael,
I think you are getting close. If you truly want only 6 fragment transitions, then you are choosing the wrong option "From filtered ion charges and types plus filtered product ions" on the Transition Settings - Library tab. This will give you the union of your 6 transitions and everything else in your heuristic transition filter.

You want to choose either "From filtered ion charges and types" (which will ignore your other filter settings) or "From filtered product ions" which will apply your entire filter and only use library peaks which meet your filter criteria. The MacCoss lab often uses an "ion 3" to "last ion" filter and the "From filtered product ions" option.

It is highly unusual to choose the option you have chosen, and I have considered removing it entirely because of its potential to cause confusion.

I am not clear, however, on how you got to where you have chromatograms extracted from only a subset of the transitions in your Targets list. It seems like maybe you had different targets when you last re-imported your data files.

Anyway, you are close. Try one of the two options I have suggested to prove to yourself that you can get a set of targets with 6 product ion transitions matching your library and properly including neutral loss ions. Then and only then, re-import your data files, and you should be set to do your analysis.

Thanks for the new screenshots. Good luck with your research.


michael plank responded:  2019-01-21

Hi Brendan,

thanks again!
Maybe Im getting closer, but still not there yet. I changed the Transition Settings: library to 'from filtered product ions' (sc10). The peptide weve focused on so far seems correct: 6 fragments are extracted which appear to be the most intense in the library spectrum and match the filter settings (sc11; not yet re-extracted).
But what about the peptide immediately above? Only two fragments are selected, even though there are more fragment ions in the library spectrum that match the filter (b, y; ion 2 to last ion -1; sc12).


Brendan MacLean responded:  2019-01-21

Can you post your file (use File > Share - Complete to create a ZIP file) to our drop point?

Are you positive you have not manually edited the transitions of this precursor? What do you see if you hover over the precursor and then click the drop-arrow that appears to the right of the precursor text?

michael plank responded:  2019-01-22


.zip is uploaded.
I dont remember if I may have manually selected transitions on this. But the wand icon is not crossed out. That means I didnt, doesn`t it?

Thank you.

michael plank responded:  2019-01-22

Hi again,

I also went one step further and did the chromatogram extraction with the settings above (using the library; 6 ions from filtered product ions). Even on the precursor that appears correct in the target list (using the 6 top ranked fragment ions) again only the two y-ions are extracted (sc14).
I assume the only setting that determines the stringency of the extraction is in Transition Settings > Full Scan > MS/MS Filterning> Resolving Power? I double checked that this is correctly set to 60 000 at 200 m/z.
When manually checking in QualBrowser (sc15), it is clear that the abundant b-98 ions that are predominant in the library spectrum are also clearly there in the run I selected in Skyline. Note however that the b-98 ions appear charge deconvoluted in the library spectrum. The y-ions that were correctly extracted were z=1 in the original spectrum. Maybe this gives a clue as to the problem? Is it normal for Skyline to do deconvolution?

Thanks for the patience,