neutral loss of glycopeptides

neutral loss of glycopeptides kpan  2017-05-04

We are now working on targeted glycoproteomics, where the glycopeptides are measured using DIA and the resulting raw files were processed by Skyline. Unfortunately, Skyline of it's current version is not quite convenient for glycoproteomic analysis. Under beam-type CID of glycopeptides, the "glycan Y ions" (or the intact peptide core plus couple of glycan moieties, see attached slide#1) usually dominate the MS/MS spectra at the m/z region >600 Th and are good targets for detection. We attempted to include the glycan Y ions in our target list in Skyline by editing neutral loss table of N-glycosylations in the modification page of peptide settings (slide #2 left). We could indeed then find the target ion in our target list (slide #2 right). The problem is that somehow Skyline limits such neutral loss ions to be bearing the same number of charges as it's precursor does, 3+ in the shown case. But actually it is more often to see these glycan Y ions with 1+ or 2+.

So here is my question, is it possible to freely ask skyline to monitor ions with the charge state that is defined by user? Or even better if one could simply define a modification as a kind of N-glycosylation and the software would automatically include corresponding glycan Y ions as targets.

Thank you very much.

I appreciate your help and am looking forward to hearing from you.
Brendan MacLean responded:  2017-05-04
This is not currently possible, but it is a good point. It seems I was too quick to jump into implementing "neutral losses" without asking whether there might also be charged losses. In the case of fragment ions, I guess it doesn't matter that much, because we allow for reduced charge states based on the fragmentation, and we have years ago now added the ability to measure the charged losses as fragments themselves. You could use the Transition Settings - Filter tab to define "Special ions" that would measure these.

Hopefully, this is the last missing piece to supporting everything. I suppose we could change the label in the Edit Structural Modification form from "Neutral losses" to just "Losses", and then add a field in the Edit Loss form that allows you to add a list of allowable charge losses for the loss, with blank meaning only neutral loss is allowed. Then when we calculate these precursors with losses we could also enumerate through the charge losses. I guess this would also allow us to simply add the charged losses as ion options as well without requiring the use of "Special ions".

Does that sound like it would meet your needs?

Once I have a clear idea of how to implement a feature in Skyline that meets your needs in a way that is coherent with the rest of the software, I think we can probably get this implemented relatively quickly.

Thanks for your feedback.

kpan responded:  2017-05-04
yes, as a workaround the "Special ions" option works. But since the masses of glycan Y ions vary according to their peptide sequence, one has then to edit individual "Special ions" one-by-one manually. It unfortunately involves lots of continuous mouse clicking.

Your solution sounds like a good plan. Just one more thing to add. Instead of editing each glycan structure one-by-one in the modification page, would it be possible to configure many modifications in a batch processing manner, such as importing a pre-defined table to Skyline? In glycoproteomics, the number of all possible glycan structures can be easily over 50 and it would be great if one could include all the possible forms, together with corresponding calculated "losses", in a table and load the table directly to skyline.

Thanks for your help!
Brendan MacLean responded:  2017-05-04
I am afraid that last part is getting a bit out of scope for a quick fix. I will note, however, that we have a grant proposal submitted with David Muddiman for improving Skyline for glycoproteomic applications. If that gets funded, then improving support beyond smallish tweaks will become a priority for us in the coming years.

Thanks for your feedback. And thanks for confirming that expanding the neutral loss support just a little to include charged losses will remove the blocking issue of not being able to measure charged losses from the precursor.

Brendan MacLean responded:  2017-05-04
I have posted a new issue to the issues list for the and linked it to this thread for supporting charged losses from precursors:

I also realized that the modifications are stored in XML in both the Skyline document (.sky file) and saved settings (*.skys file). If you have any programmers around, it should be pretty easy to have them generate a .skys file for you that contains all the modifications you are interested in. Of course, once you get this into a single Skyilne template document, then every instance of Skyline that opens the document will get all of the modifications, and you can just pass around the template as a starting point, or again use a .skys file, which you can produce by using Settings > Save Current and then Settings > Share.

Open the .skys file in a text editor and you will see a section like this:

      <peptide_modifications max_variable_mods="3" max_neutral_losses="1">
          <static_modification name="PAB-M" aminoacid="M" formula="C8H7O">
            <potential_loss formula="C9H10OS" massdiff_monoisotopic="166.045236" massdiff_average="166.24145" inclusion="Always" />
          <static_modification name="Methylthio (C)" aminoacid="C" variable="true" formula="H2CS" unimod_id="39" short_name="MSH" />
          <static_modification name="Deamidated (NQ)" aminoacid="N, Q" variable="true" formula="O - HN" unimod_id="7" short_name="Dea" />
        <heavy_modifications />

Just have a programmer take whatever tabular format you were hoping we would support and convert it to this XML or do it yourself with find and replace in a text editor.

That's about the best I can offer for now. Hope it is enough to keep you moving on your research using Skyline.

Again, thanks for your feedback.

kpan responded:  2017-05-05
Thank you very for your help! Looking forward to the fixed version.