Prosit in skyline

support
Prosit in skyline Tobi  2019-12-05 02:54
 

Dear all,

thank you very much for integrating Prosit into Skyline. I noticed that the dotp in the new library match mirror view does not treat y ions from deferentially labeled peptides as equivalent. Is it possible to extract the Comparison dotp via results grid for all peptides at once?

Best,
tobi

 
 
Tobi responded:  2019-12-06 02:29

Edit:

Being able to quickly adapt the Prosit NCE in the Library mirror view is great, having this available for the dotp between Prosit library and measurement would be nice (for example under Tools/Options/Prosit when changing the NCE there I would expect to see a change in dotp for the data or in the target list.

Also the issue with isotope-labeled y-ions can be seen everywhere, meaning Prosit in Skyline can only be used for light peptides as of now.

In total this is still a powerful imporovement, thank you very much for all the great features being added, it means a lot.

Best,
tobi

 
Nick Shulman responded:  2019-12-06 10:42

Tobi,

Can you send me your Skyline document? I probably can reproduce this problem without that, but it will make things easier for me.

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 and spectral libraries.

If that .zip file is less than 50MB you can attach it to this support request.
Otherwise, you can upload it here:
https://skyline.ms/files.url

Thanks!
-- Nick

 
Tobi responded:  2019-12-07 02:44

Dear Nick,

thanks for the swift response and sry I forgot that, its quite small so see attached, hope it contains everything.

Best,
tobi

 
Nick Shulman responded:  2019-12-09 12:37
tobi,

Thank you for sending that file.

The problem is that when Skyline creates a spectral library from Prosit, Skyline uses the spectra for the light versions of the peptides, but writes the heavy version of the Peptide Modified Sequence to the .blib file. Therefore, when Skyline displays the spectrum to you in the Libary Match window, Skyline thinks that it's supposed to be the heavy version of the spectrum. For this reason, the masses of all of the y ions are incorrect by a few daltons, so none of the y ions get annotated properly.

I will fix this so that when you create a .blib file the spectra have the correct modified sequences.

I hope this will fix the problem that you reported. I am not sure that I understood everything that you have said here.
-- Nick
 
Tobi responded:  2019-12-09 12:51
Dear Nick,

thanks for the great response.

The dotp in the mirror view should consider heavy labels as well as different sets of ions ( a selfmade library might not contain y1 and y2 but prosit does and this should not reduce the dotp).

Thank you very much for the fix, its an exiting feature and quite important as it impacts retention time standards as well.


Best regards,
tobi
 
Brendan MacLean responded:  2019-12-09 12:51
That should definitely help. Yeah, that is a bug. It took some explaining (to Tobi Rohde who implemented the features) that we needed to communicate the structurally modified peptide to Prosit without the isotope labeling modifications (because Prosit doesn't understand those and they don't impact fragmentation) and that Skyline was already built to match spectra assigned to the light modified sequence to the heavy precursor inside of Skyline.

But, if the library contains a spectrum for the light modified sequence labeled with the heavy sequence, then, as Nick notes and you have seen, the translation won't work correctly.

Thanks for reporting the issue and providing files.

--Brendan
 
Tobi responded:  2020-03-17 09:35
Hey all,

thank you for quickly implementing and fixing Prosit. It works really well with one small exception I want to highlight here again because its important and not to forget.

Please find attached. Applies to current Skyline and Sklyline daily as of 2020-03-17

1) The prosit mirror is still containing label shifts if the peptide was labeled (needs to match the state of what is compared to prosit mirror)
2) when matching a library with heavy peptide entries against heavy prosit mirror the y-ions are not matched and dotp is false
3) suggestion: Can data be matched directly against prosit mirrot without the need of being extra exported as library?

Why is it important? The library match window offers a fast way to determine prosit optimal collision energies quickly in skyline. Simply load fragment data of abundant peptides, File/export/spectral library , import this library, view library match, enable prosit mirror, scroll through CE with arrow keys to find max dotp and respective optimal prosit CE.

Since many standard and retention time peptides are heavy labeled and fixing these issues would enable this nice trick.

Looking forward to see your opinion on this.

With bets wishes, tobi
 
Brendan MacLean responded:  2020-03-17 13:42
Thanks, Tobi. Definitely lost in the shuffle.

1&2) I have a fix, which I will get into our release branch for the next patch (yes, I think there might be a second) and also into our main development branch for release with the next Skyline-daily.
3) Is really a big performance issue having the spectra cached locally versus always querying them remotely with a server running the predictions every time. I suspect that people with slower connections would be very unhappy, working off-line would become impossible, and likely the Kuster lab would not appreciate us moving to always relying on their server for everything we do with library spectra without any local storage.

So, no plan to change the overall model of how Skyline and Prosit integrate right now, but I can get you a fix for the display issues pretty quickly (in Skyline-daily). Sorry it took so long for such a simple bug. Thanks for the clear explanation which made it easy to pinpoint.

--Brendan
 
Tobi responded:  2020-03-18 03:35
Dear Brendan,

thank you for the swift response and bug-fixing willingness, I will test it throughout and report back how it looks.

For 3) I just want to say I meant it in a minimalist fashion, when no library is selected and library match window is up that this window picks a spectrum representative just for the single peptide and file selected live. The work for Prosit mirror would be the same in the workaround of exporting and reimporting the whole data as a single spectral library, but I have no idea about overall performance which would be nice to keep as you said nicely. I do not aim to have the live overall prediction for everything in the document grid, that would be too much and too unfocused.

The idea is to relief Prosit from work, when doing CE optimization in Skyline you dont need to feed Prosit with GB of raw and msms files anymore and it will be universal even for other instrument and file types.

Best regards,
tobi
 
Brendan MacLean responded:  2020-03-23 07:21
Hi Tobi,
I am aiming to release a Skyline-daily today with these fixes. On #3, I am now just confused, because I thought what you describe is what we do when you right-click the Library Match view and make sure Prosit is checked. Then Skyline should show you the predictions live on the fly, just as it does when you have a library and right-click and make sure that Mirror is checked. You seem to have found the Prosit and Mirror options in the right-click menu. So, I am not clear on what you are asking for. I am still not getting anything new we should implement out of #3. Sorry. Don't get the reference to CE optimization either. Try being super explicit and maybe providing a screenshot mock-up of what you are suggesting.

Thanks for your feedback.

--Brendan
 
Tobi responded:  2020-03-23 11:12
Dear Brendan,

thanks for the responsiveness.

The attached file shows data but no selected library. The suggestion was instead of showing just the "no spectral library match" to pick the representative spectrum from the selected peptide and raw file for the upper half of the mirror view. The workaround would be to export the data as library and import this library.

Since there is the doable workaround its not a major issue, but it would make Skyline the best default option for Prosit CE optimization for all kinds of files and instruments exceeding the Thermo raw file and msms limitation of the Prosit website lowering their load.

With best wishes,
tobi