Bug in document grid reporting "Identified" where no ID jmeyer  2020-07-08 09:22
 

Dear Skyline team,

I believe I found a bug in skyline reporting identifications in runs where the peptide was in fact not identified.

Please see the two attached screenshots where there is clearly no identification line in the chromatogram, but the document grid reports "TRUE" for an identification in that run (top line in the document grid). These are two separate peptides. When I manually drag the chromatogram to re-integrate anywhere in that run, even the same area, it goes to FALSE.

Example 1 is R.SEQEDEVLLVSSSR.Y [27, 40] ++

Example 2 is R.YPDQWIVPGGGMEPEEEPGGAAVR.E [41, 64] +++

I also attached a minimal skyline share with these two peptides, and the other peptides in the same protein.

Any help fixing this or providing a workaround would be greatly appreciated.

Best regards,
Jesse

 
 
Nick Shulman responded:  2020-07-08 10:02
Title: Bug in document grid reporting
Jesse,

The .sky.zip file that you attached seems to be missing the .skyd file (which contains the extracted chromatograms).
Can you try creating that .sky.zip again and send it to us?

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

By the way, the "Identified" column in the document grid tells you whether there was an MS/MS identification between the start and end times of the chosen peak. So, we would definitely expect that if you moved to peak to some other part of the run where there were no IDs, that would change to "FALSE".

Skyline shows the location of the MS/MS IDs on the chromatogram in two ways. Identifications that happened in the file that the chromatogram came from are shown in dark blue, with the word "ID" at the top. Your screenshots do not show any of those matching ID's, but that might be because you need to right-click on the chromatogram and turn on "Peptide ID Times > Matching".
Skyline also shows MS/MS IDs that happened in other runs, but which were mapped to the times in the chromatogram's run using retention time alignment. Those aligned ID times are shown as light blue lines like in your screenshot.

The "Identified" column in the chromatogram tells you whether there was an MS/MS ID within the integration boundaries, and it will be either "TRUE", "ALIGNED" or "FALSE".
I do not believe that there is an easy way to find out whether the peptide was identified anywhere along the chromatogram, regardless of the chosen integration boundaries.
BiblioSpec spectral libraries (.blib) files are SQLite databases, so, if you are familiar with database tools, you could query this information by JOINing the RefSpectra, RetentionTimes, and SpectrumSourceFiles tables.
-- Nick
 
jmeyer responded:  2020-07-08 12:18
Thanks Nick,

Separate issue - the only way to get the share to have a skyd file was to open the share, wait for it to re-extract chromatograms, and then share that. Find that share attached. This document was saved from another document if that helps explain that.

I meant even if I manually update the extraction to be the exact same boundaries as originally, the incorrect "TRUE" value goes away in the document grid.

As you'll see in the share "Peptide ID Times > Matching " is on, and there are no IDs in the integration boundaries.

Since it seems the most MS/MS IDs are within the integration bounds (other than these false positives) I'm okay with assuming this is a good proxy for whether the peptide was IDed.

Thanks for your help,
Jesse
 
Nick Shulman responded:  2020-07-08 12:53
Thanks for sending this new .sky.zip file.

The bug is that the chromatogram graph is only displaying the ID lines for MS/MS IDs that have the same charge state as the selected precursor.
The charge state of a peptide does not affect its retention time, so Skyline should be displaying all of those matches with different charge states as real "ID" lines.
So, in your examples, the Precursor Result's "Identified" value really is supposed to be "TRUE", but the chromatogram graph incorrectly makes it look like there are no matching IDs.

When you change the integration bounds on the chromatogram graph, Skyline uses the same faulty list of ID times which does not include the other charge states, and that's why the "Identified" switches to "FALSE".

It looks like this behavior of not displaying other charge states' ID times on the chromatogram graph has always been like this.

Thanks for reporting this bug. We will try to fix it soon.
-- Nick
 
jmeyer responded:  2020-07-08 13:11
Thanks Nick,

Thanks for clarifying and for figuring out the problem so quickly.

It sounds like the solution could go 1 of 2 ways here, and I think it really should only show lines/ID=TRUE for the selected precursor. I think it is not intuitive and misleading if Skyline says that any charge state was IDed when you select a particular charge state.

In the example I sent, each of these runs are different FAIMS compensation voltages, so we want to know in which runs that charge state was identified (and therefore which FAIMS CV can be used to look for that charge state).

I can also imagine using different mass spec source or ionization settings and using skyline to look when a particular charge state was IDed, which would also be confusing if any ID was called "ID".

I suppose a third option would be to have a second option for the table that was something like "IDed in this run" that follows different rules, as long as the rules are explained when you hover over the options as you have done so nicely for most of the options.

Thanks again, and please ping me when you have a chance to fix this one.

Best regards,
Jesse