No spectra were found for the new library

support
No spectra were found for the new library joshuasmith  2021-12-03
 
Hi Skyline team,

I know that this issue has been the subject of several previous support tickets, but after reading through all of them and trying to figure out what is happening with my data, I am still having an issue with building a spectral library in Skyline using my DIA search results.

I have overlapping-window DIA data from a Fusion Lumos. I know the raw data isn't an issue, as I can import the .raw files into Skyline with a target list from a previous experiment just fine (see "DIAimported" screenshot). And I get valid .tsv output data from MSFragger. I used FragPipe for my search, first doing DIA-Umpire signal extraction on mzXML converted files, then searching the pseudo-DDA data with MSFragger in Open Search mode.

I have my interact.pep XML file in the same directory as the calibrated .mgf, .mzML, and .pepxml MSFragger output files.

I opened Skyline, saved my document, and navigated to File->Import->Peptide Search. I attempted to build a new library by adding my interact.pep search result file, with the DIA workflow selected. I hit Next, and then for a hot second, it looks like it is reading in the spectra from the .mzML files that are in the same directory as the interact.pep search result file (see "LibraryBuild" screenshot). But then it immediately returns the error below (also see "LibraryBuildError" screenshot).

Not sure what I'm doing wrong, any help would be appreciated. I have attached screenshots, my XML search result file, and the FragPipe/MSFragger log file.


Thanks,
Josh

---------------------------
Skyline
---------------------------
ERROR: No spectra were found for the new library.

Command-line: C:\Users\jossmith\AppData\Local\Apps\2.0\TZE299RM.CXM\MA2X4L6X.993\skyl..tion_e4141a2a22107248_0015.0001_62a9d240a069a6b4\BlibBuild -s -A -H -o -c 0.95 -i PAA3 -S "C:\Users\jossmith\AppData\Local\Temp\tmp5A85.tmp" "C:\Users\jossmith\Documents\Adductomics\ColeLabData\211109\PAA\DIANN\PAA3.redundant.blib"
Working directory: C:\Users\jossmith\Documents\Adductomics\ColeLabData\211109\PAA\MSFragger
---------------------------
OK More Info
---------------------------
System.IO.IOException: ERROR: No spectra were found for the new library.

Command-line: C:\Users\jossmith\AppData\Local\Apps\2.0\TZE299RM.CXM\MA2X4L6X.993\skyl..tion_e4141a2a22107248_0015.0001_62a9d240a069a6b4\BlibBuild -s -A -H -o -c 0.95 -i PAA3 -S "C:\Users\jossmith\AppData\Local\Temp\tmp5A85.tmp" "C:\Users\jossmith\Documents\Adductomics\ColeLabData\211109\PAA\DIANN\PAA3.redundant.blib"
Working directory: C:\Users\jossmith\Documents\Adductomics\ColeLabData\211109\PAA\MSFragger
   at pwiz.Common.SystemUtil.ProcessRunner.Run(ProcessStartInfo psi, String stdin, IProgressMonitor progress, IProgressStatus& status, TextWriter writer, ProcessPriorityClass priorityClass) in C:\proj\skyline_21_1_x64\pwiz_tools\Shared\Common\SystemUtil\ProcessRunner.cs:line 142
   at pwiz.BiblioSpec.BlibBuild.BuildLibrary(LibraryBuildAction libraryBuildAction, IProgressMonitor progressMonitor, IProgressStatus& status, String& commandArgs, String& messageLog, String[]& ambiguous) in C:\proj\skyline_21_1_x64\pwiz_tools\Shared\BiblioSpec\BlibBuild.cs:line 201
   at pwiz.Skyline.Model.Lib.BiblioSpecLiteBuilder.BuildLibrary(IProgressMonitor progress) in C:\proj\skyline_21_1_x64\pwiz_tools\Skyline\Model\Lib\BiblioSpecLiteBuilder.cs:line 157
---------------------------
 
 
Nick Shulman responded:  2021-12-03
There are a couple of reasons that you might get the "No spectra were found for the new library" message.
One would be that none of the peptide spectrum matches had a score which was above the cutoff score specified on the first page of the Import Peptide Search wizard.
The other common reason is that the spectrum identifiers in the .pep.xml file are different than what BiblioSpec needs them to be. I am not the expert on what spectrum identifiers need to look like for different types of search results.
I think there are lots of other reasons that this could happen, too.

If you send us your .mzML files, we can certainly try to reproduce this error and tell you what is happening.
You can upload those .mzML files here:
https://skyline.ms/files.url

I realize those .mzML files are probably large and it will take a long time for you to upload them. We have plenty of bandwidth and storage space, so do not worry about that on our account. If you happened to have a .pep.xml file which only referenced one .mzML file, that might be easier for you to upload.
(Or, if you want to point us at some other place that we can download the files from, that works too).

-- Nick
 
joshuasmith responded:  2021-12-03
Nick,

Thanks. I am not sure it's the score cutoff issue, as I tried values from 0.01 to 0.99 with no luck. The PSMs output file I get from MSFragger (attached) has hyperscores as well as expectation probability scores. Not sure which ones Skyline would look at, although given the upper bound of 1 for the cutoff in Skyline, it seems the expect score would make more sense.

As far as the library headers, I don't know what BiblioSpec needs, but I've attached the library TSV file that I get from MSFragger - perhaps you could assess whether it is appropriate for BiblioSpec to read.

Finally, I've uploaded a whole folder of my MSFragger input and output files, including the pepxml, mzML, mzXML, and mgf files. The zipped folder is named MSFragger.zip with a description of "SmithJ_LibraryBuildError_120321".

Thanks,
Josh
 
Nick Shulman responded:  2021-12-03
Josh,

I think if you try this in Skyline-Daily, everything will work for you.
This "No spectra were found" message only happens with Skyline 21.1.
In a few weeks, we are going to release Skyline 21.2, which will also correctly work with this data.

You can install Skyline-Daily from here:
https://skyline.ms/project/home/software/Skyline/daily/begin.view?

-- Nick
 
joshuasmith responded:  2021-12-03
Thanks, that worked great! Any idea what was causing it?

Josh
 
Nick Shulman responded:  2021-12-03
I am not sure, but I think the change which fixed this might be that when the .pepXML file comes from MsFragger, BiblioSpec now first looks for the "_calibrated.mgf" file before it tries looking at the ".mzML" file.
The spectrum identifiers in the pepXML only make sense inside of the mgf file.

Here is the github page where that change was made:
https://github.com/ProteoWizard/pwiz/pull/1752
-- Nick
 
joshuasmith responded:  2021-12-03
Sorry, one last thing, and this is not related - since I am using Skyline Daily, I decided to try out the Synchronize Integration feature. I love the idea, as I am often doing large batches of samples, and this would be a true life saver. However, I'm not sure what I'm doing wrong (i have an iRT calculator that i generated from the data, and I've aligned all my replicates to the iRT score), and it's not working too well - my peak integration in one replicate is not hitting the same peak in another iRT-aligned replicate.

In any case, I decided to not mess with it anymore and tried to turn it off - but I can't. I right click a replicate chromatogram and try to un-check the "Synchronize Integration" option. The dialog window comes up, I de-select all my replicates, set it to "align to none", and hit ok. But when I integrate a peak, I am still synchronized. When I right click in a replicate chromatogram again, Synchronize Integration is still marked as active (checked) in the right-click menu that appears.

What am I missing? How do I turn it off?

Thanks,
Josh
 
Nick Shulman responded:  2021-12-03
Josh,

It appears that Skyline has a bug in that you cannot change the setting from "synchronize all of the replicates" to "synchronize none of the replicates".

But, other changes of settings do work.

So, what you need to do is click on "Synchronize Integration" and choose Group By "Replicates" and make it so that one and only one replicate is checked.
Then, you can do "Synchronize Integration" again, and make it so that zero replicates are checked.
If you do that, then Synchronize Integration will be turned off.

(Also, I believe if you really do have only one replicate selected, that is equivalent to having synchronize integration turned off)
-- Nick