MS1 XIC QUAN without MS2 Library

support
MS1 XIC QUAN without MS2 Library victor nesati  2018-05-17 16:20
 

We are trying to configure Skyline to do MS1 XIC based quantitation based on precursor masses alone without prior results of database search.

We were able to create sequences of desired peptides with custom modifications in the Skyline. But getting Skyline understand that we want to use precursor masses of those peptides as "pseudo-library" for consequent QUAN was a challenge. Actually raw file was not loaded. We found work-about through building SSL file then converting it into blib, importing those library along with mzXML file, and then getting desired info, but this still requires presence of fragmentation in data acquisition, which we dont really need, as we do know the masses we are looking for and put it simply we are just trying to use Skyline instead XCalibur to get peak Area data. Any suggestions on how to handle this situation ?

Interestingly enough I also found that Panorama could be used to create Chromatographic Libraries which might be applicable to what we are trying to do. In addition I found Topograph software which from the looks of it might be able calculate chromatographic peak areas extracted from MS1 scans in DDA data, which sounds exactly what we are trying to do.

I would appreciate your pointing me into right direction.

Best

Victor

 
 
Nick Shulman responded:  2018-05-17 19:15
I am not sure that I understand your question.

If you want to tell Skyline to look for particular m/z values without regard to what sort of peptide sequence they might correspond to, then you probably want to tell Skyline that these are small molecules.
The small molecule tutorial might be helpful:
https://skyline.ms/wiki/home/software/Skyline/page.view?name=tutorial_small_molecule

However, it sounds like you were able to tell Skyline about peptides that had the right m/z values, but something else went wrong when you tried to extract chromatograms.

The MS1 full scan filtering tutorial sounds like exactly what you are trying to do:
https://skyline.ms/wiki/home/software/Skyline/page.view?name=tutorial_ms1_filtering

But, if that does not help, then maybe you should send us your files.

In Skyline, you can use the menu item:
File > Share > (complete)
to create a .zip file containing your Skyline document and supporting files including extracted chromatograms.

You can upload that .zip file here:
https://skyline.ms/files.url

You should also send us at least one of your .raw files and we can figure out what is going wrong.
 
Brendan MacLean responded:  2018-05-17 19:31
Agreed. Have a look at the MS1 filtering tutorial and the two webinars referenced at the bottom of that page. These are all about extracting MS1 signal from DDA data. If you are trying to do something different from what they describe, please explain.

—Brendan
 
victor nesati responded:  2018-05-17 20:53
Hi guys,

We explored MS1 tutorial to death, already using it in day-to-day operations without any problem so we are way way way over it. What we want is to skip database search ALLTOGETHER and implement XIC quan assay based on precursor masses alone . Those masses were IDed manually
and thats why we just need to define those masses as ID parameters one way or another, and let Skyline use those precursor masses
as basis for XIC Quan. We tried to implement " Import Peptide List" workflow. We defined our peptides all-right, and they became visible in the left panel. Though we could NOT define them as library, we could only store this "project" as file with .sky extension. What remained not clear on how to make a jump from defining those peptides as sequences to telling Skyline to use them as library . Is it possible at all ? Additional complication appeared when we tried to import mzXML after manually defining our peptide list in an attempt to use it as library. For some reason raw file import was not successfull.

Anyway, suggestion to define our "peptides" as small molecules is very interesting one and I will take a look at the corresponding manual.
Dont know how it handles "library" but if it is possible to just use precursor masses alone for any further Skyline operations that would be very very fine with me and indeed cover what we want to do.

I will let you know how it went along.

Cheers

Victor
 
Nick Shulman responded:  2018-05-17 21:17
I am confused.

Why do you want to create a library? Libraries are not necessary for MS1 chromatogram extraction.

You can get peptides into your document by doing:
Edit > Insert > Peptides
and you can specify amino acid modifications in the dialog by their masses (e.g. "PEPTIDEC[+57.02]K").

A spectral library does give Skyline hints about what retention time to expect a peptide at, but there are other ways to accomplish that such as, I think, specifying the "Explicit Retention Time" for the peptide in the Document Grid.
A spectral library also tells Skyline which MS2 transitions are expected to be present, and their relative intensities.
 
victor nesati responded:  2018-05-17 22:00
Hmm, for now we only managed to perform XIC MS1 QUAN when having "proper" library resulted from database search.
May be there is another way described in small molecule memo.
We will try to implement suggested Peptide import workflow and will include detail PPT if something will go wrong.
Cheers
 
Brendan MacLean responded:  2018-05-18 16:10
I get the sense that there is some confusion here around the necessity of a "library" for targeted extraction of XIC peak areas from mass spec data. Perhaps this is due to how frequently "Spectral Library" and "Assay Library" get used in relation to DIA data processing.

In the common use of "Library", "Assay Library", etc. emanating from Zurich, they are referring to:
1. A set of XIC coordinates in terms of m/z values, typically called "transitions" in SRM.
2. Normalized retention time values for each analyte.
3. Expected relative ion abundance for each transition in the set.

I have at times referred to this as an "Enhanced Transition List" and you can still import one into Skyline through File > Import > Transition List, but we have also added File > Import > Assay Library due to growing use of this tabular format in the proteomics community.

Skyline supports these functions with:
1. The Targets view and the transitions in it.
2. An iRT library, which can now be embedded in the spectral library if enough information was available at spectral library build time.
3. A spectral library typically built from search results.

For XIC from MS1-only, as Nick noted, you definitely do not need #3, we can calculate the expected relative ion abundance of the precursor isotopes you target without needing any empirical measurement.

The most common MS1-only extraction is definitely from DDA data where there are also peptide spectrum matches (PSMs) to the data-dependent MS/MS spectra, which you clearly understand from the MS1 Filtering tutorial. But it seems that you may not have these for your data files? Are your files not actually DDA, but MS1-only? Or you are trying to target precursors in MS1 that were not identified in the MS/MS spectra?

If you have PSMs, even partial coverage, then they are certainly your best source of #2 ("normalized" retention times), and Skyline will align and translate ID times between runs to know where to look in the MS1 chromatograms for your peak of interest. This would be the only reason I can imagine for you to build a spectral library in Skyline to get ID times in the runs you are extracting from into Skyline and allow for run-to-run alignment. Here is a Tip page on dealing with missing ID times building spectral libraries from searched DDA data:

https://skyline.ms/wiki/home/software/Skyline/page.view?name=mascot_missing_rt

Working with MS1 data without ID times is a bit more error-prone.

So, do you not have IDs and ID times? Using Explicit Retention Time will be less forgiving because there will be no alignment. Skyline will just favor peaks close to the explicit time in minutes that you set.

If you have no IDs, you can still build an iRT library and use iRT for retention time prediction as described in the iRT tutorial.

Inside Skyline for MS1 XIC, you should be thinking about your targets (in the Targets view) and what you will do for retention time prediction where you have these 3 options in order of preference:
1. Times in the files being processed corresponding to MS/MS scan times which were matched to your target. (ID times)
2. Normalized retention times (iRT values)
3. Explicit retention times, when you have high confidence and low variance in your chromatography.

I guess I will stop there. If nothing I have said makes this any clearer, then probably we should arrange a web meeting to talk about this and perhaps you can give a bit more information, or show an example, if you can't send files or post screenshots to clarify.

Hopefully we can help you achieve what you are striving for.

--Brendan
 
victor nesati responded:  2018-05-20 16:06
Hi Brendan,

Thanks for very clear explanations. I think part of the confusion was due to very simple fact that I did not press " SAVE " button while inserting either PEPTIDES or SMALL MOLECULES. Once this little trick was mastered everything else fell in place. And I was able to do MS1 XIC based quan based on formula or peptide sequence alone without inserting results of prior search ( library). So in a nutshell my "library" was set of m/z values calculated by Skyline from peptide sequence or chemical formula. That was sufficient for our purposes.

The necessity of using this workflow is two fold. For example ID of glycopeptides is really an art , quite frequently it is done by hand so all you have is RT, chemical formula, oxonium ions, couple of other signature fragments and thats about it. So there is a need to cut the chase and go straight to the beef using m/z traces alone. The same goes to MS1 based XIC quantitation of cross-linked peptides. It is much easier for me to specify their formula and work with them as small molecules rather than go through other elaborated schemes. Though now as I mastered an art of converting custom created SSL to blib that certainly provides another avenue. So DIA or DDA are definitely part of ID workflow, but once you know what you are dealing with we feel that use of full scans alone without any fragmentation is a better way for relative quantification we are gunning for. Bear in mind that we are working in protein characterisation environment which is quite different from standard proteomics workflows geared for complex mixtures.

Couple of silly technical issues. When I am working with small molecules I also want to specify number of isotopologues (isotopic traces) used for quantitation. There is no specific button for that in the small molecule transition settings. What I am doing is specifying number of isotopologues (ion charges) for peptides and hoping that they will be applicable for small molecules. Sometimes it works and sometime it does not. Quite frequently upon import of raw file I have only major isotope visible despite specifying 3 ions charges and doing some silly thing in those transition settings brings desired number of ion charges. Would be nice to learn on how to do it consistently.

LAst one. Sometimes a mistake is made in chemical formula, which is discovered from missing a particular m/z trace in already imported raw file. The usual way to fix it is through hitting " MODIFY" button for this particular trace and insert right formula. However, for some reason this new trace with newly corrected is not updated automatically even upon re-import of raw file and the only way to fix this mistake is to repeat the whole procedure from the scratch. Is it true or there is another way to re-process same raw file using corrected version of chemical formula.

Cheers

Victor
 
Brian Pratt responded:  2018-05-21 09:54
Hi Victor,

>> There is no specific button for that in the small molecule transition settings

It's true that we still have work to do to fully separate the peptide oriented UI from the small molecule UI, or in this case to make it clear which parts are applicable to both. Although in this case the setting is in the Full Scan tab, which is free of peptide-specific language, so I may not be understanding your point.

>> What I am doing is specifying number of isotopologues (ion charges)

I think there's some confusion here - ion charge is determined by the adduct. Isotopologue count is set in the Full Scan settings MS1 Filtering "Isotope peaks included" control.

That said, it's true that you have to go find that setting and make some kind of change to provoke Skyline into performing the calculations and expanding the targets list to include the new precursors. So far with small molecules our users have generally wanted their imported transition lists to be treated quite literally, so we've not done that by default. Given that Skyline doesn't know how to predict fragmentation for generalized molecules as it does for peptides, this makes some sense. But for your workflow we should make this easier to achieve - we will keep this in mind as we continue to make UI improvements for Skyline in small molecule usage.

>> or some reason this new trace with newly corrected is not updated automatically even upon re-import of raw file

I haven't experienced that - perhaps you can provide a data set to demonstrate? Use Skyline's File>Share>Complete to create a .sky.zip file, and upload that along with the raw file to http://skyline.ms/files.url .

Thanks,

Brian Pratt
 
Brian Pratt responded:  2018-05-21 10:55
Hi Victor,

Apologies, on rereading your comments I see that I misunderstood: you are looking not for isotopic distributions, but rather isotopic labeling for quant. That is indeed specified in the adduct (though it's still not about charge per se), as described at https://skyline.ms/wiki/home/software/Skyline/page.view?name=adduct_descriptions.

If I can see the transition list you're starting with, I may be able to make some useful suggestions for quickly getting to the Skyline document you want.

Best Regards,

Brian
 
victor nesati responded:  2018-05-21 15:35
Hi Brian,

Thanks for clarifications. Indeed there was confusion between "ion charges", which looks like a definition of different adducts ( good to know) and isotope peaks , whose number is defined in Peaks Insert Space. What I meant in my previous message was indeed the number isotope peaks ( isotopologues) and this point is now clear.

Though let me go back to Filter Tab in Transition Settings. If I understood things correctly for Peptides

Precursor charges refer to number of charge states I wish Skyline to use for Peptides I inserted.
Ion Charges is the number of adducts. In case I am M+H freak I put those number of charges to 1 and live happily thereafter , correct ?
Ion Types: p for precursor and if I have fragments y,b etc

When I am switching to Small MOlecules Tab I have

Precursor adducts . In cased I specify my peptides as small molecules and want to make sure that Skyline takes into account +1, +2 etc charge states I put my precursor adducts as [M+H], [M+2H], [M+3H] etc, right ? For metabolomics probably adding Na and K adducts might be helpfull as well.
Fragment adducts. Probably the same idea except may be adding water and ammonia losses might be usefull.
Ion Types: Same as in Peptides

Couple of other points. We have Ion Match tolerance Tab in Transition settings and mass accuracy settings in Full Scan Tab. Could you pls explain what refers to what, and what is applicable for general DDA workflow and what is for simple MS1 XIC Quan ?

Cheers

Victor
 
Brian Pratt responded:  2018-05-21 16:47
In that case if the Filter tab we have already gone to some effort to split out the UI for peptides vs small molecules. Anything on that peptide tab is ignored for small molecule transitions.

When working with peptides, we can speak in terms of "charge" with the understanding that we mean (de)protonation. So on that Peptides tab you can say "1,2,-4" and internally Skyline uses adducts [M+H], [M+2H], and [M-4H]. That's a convention proteomics users expect. While we could theoretically support sodiated peptides, we don't because the peptide fragmentation model assumes protonation.

When working with generalized molecules, we have to speak in terms of adducts because we need to support more ionization modes. As you have observed, there are many ways to attain any given ion charge, so we drop the charge shorthand and go with full adduct specifications.

Ion types is another major difference between those two tabs - for peptides, we can perform insilico fragmentation so we allow you to specify which kind of predicted fragments to include in the Targets window. For generalized molecules, we don't have that fragmentation prediction capability so we just have "p" and "f" where "f" is a catch-all for "fragments". The only fragments Skyline knows about are found in the transition lists you import, and there's currently no sense of fragment type per se.

Because of our inability to do small molecule fragment ion prediction, Skyline is pretty shy about automatically updating the Targets list based on the Filter tab, and unlike with peptides we don't actually use that list of adducts to generate anything, it's mostly used for filtering Library Explorer etc. In practice, it's necessary to specify the various adducts you want in the transition list at import time. The Ion Types control does matter, though, and folks get tripped up on that when f and/or p is or isn't enabled. (I have an action item to make that set itself properly based on transition list import contents)

This thread may be useful in understanding Ion Match tolerance vs mass accuracy.:
https://skyline.ms/announcements/home/support/thread.view?rowId=3602

Best Regards,
Brian
 
victor nesati responded:  2018-05-21 18:38
Thanks for those explanations, Brian. Will spend a bit of time digesting mass tolerances issues.
I figured out the problems with updating of new traces.
Looks like I needed to physically delete old one, manually highlight new one and then everything
fell in place.
Last thing for now. Could you please clarify if X axis labels could be put in full view.
For now they are shortened in quite stochastic fashion.
I am enclosing PPT slide highlighting the issue.
Cheers
Victor
 
Brian Pratt responded:  2018-05-22 07:59
Have you tried adjusting the font size? Right-click on the graph, then choose Properties. Maximizing the window may help too.

I agree that we could do a better job of finding abbreviations, though.

Brian
 
Brendan MacLean responded:  2018-05-22 09:24
I will note that Skyline also decides on when to shorten the names and shrink the font for x-axis labels based on the amount of vertical space available for the main graphing region. You may be able to get longer names and larger font by increasing the height of the graph. In some cases, you may also need to force a change to the graph. I see this mostly with cases where the graphing engine has decided to drop x-axis labels because there are too many to show. Then when you widen the graph to make room, sometimes that is not enough to cause the missing labels to show up and I need to change selection or something to get the graph to refresh the x-axis labels. Wish we could figure this one out, but it is mostly a minor nuisance.

Turns out it is a pretty tricky problem preserving the graphing area enough to show something in the face of all possible x-axis labels, and when the graphing area shrinks to nothing, our graphing engine blanks the entire pane, which generates far more support issues than font changes and text reduction.

Hope you are still able to get something close to what you want. Your ultimate out is to use right-click > Copy Data to copy the data and paste it into another program for graphing (e.g. Excel or a text editor to save a file for R or something else). And then you can replace the shortened names with their full equivalents (aided by the Document Grid) and format exactly how you want if you are seeking publication or poster quality plots.

Thanks for the screenshot with your post.

--Brendan
 
victor nesati responded:  2018-05-22 22:59
I am cranking this exercise a notch and now would like to make MS1 XIC based quantitation of the RAW file, acquired using DDA and thus containing all fragment info, being conditional on the presence of specific fragments (special ions). Steps which I have taken.

1. Inserted Peptides
2. Defined their modification
3. Loaded RAW file
4. Made sure that I am monitoring right charge states and right isotopologues

So pure MS1 quan works impeccably and all traces were green.

5. Then I defined my specific special ions in the transition settings ( VT_b2, VT_b3, VT_y3) and made sure that all or one of them is ticked.

As soon as fragment ions ( even existing ones) appeared in the target list all green button turned yellow , probably specifying some uncertainty.
Not so sure why as I know for fact that those fragments are there, so why turn yellow all of a sudden ?

Second question is it possible to tell Skyline to forget about all other fragments it is seeing, and focus only on special number of ions I specified in special ion tab. Could not find this option in Filter Tab of the TRansition Settings. It is always either from precursor or ions 1-4. Though if my specific ions are located on the terminus such as y2, b2, b3 then it is possible to circumvent this issue by using Product Ion selection from IoN3 to Ion 1 or 2. However if my specific ions are located throughout peptide sequence then it is a bit of a challenge, it it ? Ideally I would like to forget about product ion selection alltogether and only focus on specific product ions and precursor traces.

I enclosed a bit of an illustration.

Cheers

Victor
 
Nick Shulman responded:  2018-05-22 23:16
I think you just need to tell Skyline to extract the chromatograms again after you have added these new transitions.

You do this by going to:
Edit > Manage Results > Reimport

When you add new transitions to your document, they will not have any chromatogram data until your reimport.

I do not know the answer to your second question.
 
victor nesati responded:  2018-05-22 23:29
Thanks NIck,
Reimport does not help. I am on this trick.
May be Brian or Brendan can wade in.
Cheers
 
Brian Pratt responded:  2018-05-23 11:10
The yellow indicates that only some targets have chromatograms. In your second slide, you can actually see a case of this about half way down the list.

>> Ideally I would like to forget about product ion selection alltogether and only focus on specific product ions and precursor traces.
This sounds like Skyline's small molecule capabilities might be a better fit if all that nice peptide-aware automation is just getting in your way.

If reimport isn't working, there is the possibility that the data for that transition really just isn't in the raw file, or of course there could be a Skyline problem. If you can provide your Skyline document (File>Share>Complete in Skyline to create a .sky.zip file) and the raw data, we could have a look. You can upload that to http://skyline.ms/files.url .

Cheers

Brian
 
victor nesati responded:  2018-05-23 14:51
Hi Brian,

I decided to use your own MS1 tutorial data to investigate this issue a bit deeper.
Provides much better study case and there associated fragment library allowing to see fragments.
Here are my findings ( enclosed).

1. For some reason most intense fragment signals are not marked as green in target list ( ex. y8 ) in Slide 1
2. Question arises are there any fragment signals that actually marked green ? Answer is Yes : Only 1 . X5 minus 98 (Slide 1)
3. When I remove rock solid Y8 fragment from target list leaving X5-98 Yellow button for peptide turns green. ( Slide 2)
4. Inspection of area of X5-98 ( m/z=615.2773) around 615 does not show any assignment except grey signal at 615.32953 (Slide 3)
5. My settings are in SLide 4

The only conclusion that I can make that unless my settings are fundamentally wrong there is something interesting with Skyline picking legit fragment signals in target list. Interestingly enough fragment peptide coverage ( Slide 1) clearly shows y8 highlighted in black so there is some miscommunication between what Skyline sees in Library Match and what it shows in Target list as legit signals.

My question did either you or Brendan manage to get legit fragment signals shown in target list being highlighted in green in MS1 tutorial?
If yes what version of Skyline was used and with what settings. Otherwise we can just flag it as an issue and I will try to deal with the problem through report template by incorporating output of desired fragment signals.
 
Brendan MacLean responded:  2018-05-23 15:27
Hi Victor,
You can rid yourself of this unfortunate effect by reducing your Transition Settings - Instrument - Method match tolerance m/z from 0.055 to something much smaller, like 0.001. The m/z values that Skyline assigns during extraction will be pretty nearly exactly the target m/z values, and unfortunately it looks like Skyline is somehow getting confused between the x5-98 transition (615.2773) and the precursor [M-1] transition (615.2666).

No fragment ion should match in this data, because no fragment ion extraction was performed based on the Transition Settings - Full-Scan tab having MS/MS filtering - Acquisition method = None.

This is a bug in Skyline, which I will fix. Skyline should never be matching chromatograms extracted from MS1 with fragment transitions.

It is important for you also to note that you have no reason to expect any fragment chromatograms in this data, because the settings indicate that none should be extracted.

Thanks for pointing this out.

--Brendan
 
victor nesati responded:  2018-05-23 15:57
I think we are on something here, Brendan.

I was always a bit curious regarding those MSMS filtering settings in Full-Scan Tab. From what I see we have only 3 choices
1. NONE
2. Targeted
3. DIA

As I did not see DDA than I was stuck in first gear.
Though philosophical question arises what happen with us poor souls stuck with age old DDA ?
Suppose we have DDA data which one of remaining two choices ( DIA/Targeted) should work better for highlighting fragment ions if it work at all ?

Cheers
Victor
 
Brendan MacLean responded:  2018-05-23 19:33
Hi Victor,
We generally recommend you follow the outline of the MS1 tutorial, first searching your DDA with some form of spectrum matching software tool, build a library from that and then extract on MS1 chromatograms. We have not put a lot of effort into supporting extracting fragment chromatograms from DDA data, because they are not quantitative. If you follow the intended workflow you will be able to see where matched spectra were acquired and even view them in the Library Match view.

If you choose to extract chromatograms from DDA MS/MS spectra, you are mostly on your own. Don't expect this to come out looking very good. But We have known people to do it before using the Targeted option, or even the DIA option with a very narrow window based on the precursor m/z values of the MS/MS spectra.

But, maybe just start with using Skyline as designed for MS1-only chromatogram extraction from DDA data.

--Brendan
 
victor nesati responded:  2018-05-23 20:01
Thanks Brendan,

As they say horses for courses or vice versa. :-)
Indeed Specifying Targeted Acquisition Method with a product mass analyzer set as Centroided ( thats what we have)
set all the lights green in all the right places for my own data acquired on qE. So we are all good there.

Have to tell you Skyline report builder and filtering is a thing of beauty. So simple and easy.
Just did not found yet whether there is an option to specify in the report the colour code of the specific trace ( red, yellow, green).
Should I continue looking ?

As for fragment extraction we are using them solely as qualifier thats why I put so much
attention to turning them green, while using MS1 XIC as quantifier so I believe we are all good there.

Best

Victor