Custom Spectral Library

support
Custom Spectral Library gwilson9  2018-02-13 19:58
 
Hello,

I am encountering an error building a custom spectral within Skyline. I get a series of errors after loading the .ssl file:

Error: unknown escape sequence
Error:
Error: reading file ForSpectralLibrary.ssl

The .ssl file is templated after the example found here: https://skyline.ms/wiki/home/software/BiblioSpec/page.view?name=BiblioSpec%20input%20and%20output%20file%20formats

and spectra files are .ms2 files created from Thermo .raw files with MSConvert.

I have attahed the .ssl file and one of the .ms2 files (truncated for file upload). I would really appreciate your help trouble shooting, thanks!

Gary
 
 
Nick Shulman responded:  2018-02-13 20:35
There are two things that BiblioSpec is having trouble with in your .ssl file.
One is that your lines all end with just a carriage return, but BiblioSpec expects a carriage return and a line feed.
You can use a program such as "Notepad++" to replace "\r" with "\r\n" and that will fix it (in the Replace Dialog, set "search mode" to "extended" so that "\r" means "carriage return").

The other problem is that BiblioSpec does not like the backslashes. The particular tab-separated-value parser that BiblioSpec uses thinks that backslash is a special character, so you need to replace your backslashes ("\") with double-backslashes ("\\"). Usually, people put their .ms2 files in the same directory as the .ssl, so you would not have any backslashes at all.

I have attached a fixed version of your "ForSpectralLibrary.ssl" which should work for you.
 
gwilson9 responded:  2018-02-14 07:03
Nick,

Thank you for your quick response! I am now able to upload my data using the .ssl file that you provided. However the library that is built does not contain any of the phosphorylated peptides. Is there an issue with the way that I have denoted modifications in my peptide sequences?

Thanks!

Gary
 
gwilson9 responded:  2018-02-14 08:51
Skyline also detects a number of modifications that are not in the .ssl peptide list. I'm wondering if these issues are related. Screen shot attached.
 
Nick Shulman responded:  2018-02-14 09:04
The masses of the modifications in your .ssl file are a little different than what Skyline expects them to be.
For instance, you are using "79.99" for Phosphorylation, but Skyline expects that to be "79.97" (or "80.0" would also work, as well as "79.966331").
 
gwilson9 responded:  2018-02-14 09:18
Problem solved. Thanks again for the quick response!

Gary
 
gwilson9 responded:  2018-02-14 12:30
Can I get your help once more? :)

Very few of my library spectra are annotated. Only 2 - 3 peaks per spectrum have an ion assignment at best and most have 0 assigned peaks. I have accounted for fixed and variable mods in 'Peptide Setting' and have walked through the Skyline Spectral Library Explorer Tutorial. I have also checked spectra against protein prospector's annotation tool to make sure that the IDs are valid. Is there a common mistake that you can point me towards?

Thanks,

Gary
 
Nick Shulman responded:  2018-02-14 13:11
It sounds like your .ssl file might be referring to the wrong spectra in your .ms2 file.
Does the spectrum that Skyline shows you in the Spectral Library viewer look like the spectrum where the MS2 identification happened?

It might help if you sent us your Skyline document.
In Skyline, you can use the menu item:
File > Share > (complete)
to create a .zip file containing your Skyline document and supporting files including spectral libraries.

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

You should also send us at least one of your .raw files.
 
gwilson9 responded:  2018-02-14 13:39
The MS2s do not look like the best scoring MS2s from my search. I also noticed that the ion assignments change when I assign heavy and light isotope labels, so perhaps I am doing this incorrectly.

I have loaded my sky.zip and raw files to the file share.

Thanks again,

Gary
 
Nick Shulman responded:  2018-02-14 15:04
Is there a reason that you are writing your own .ssl file, instead of just pointing BiblioSpec at your peptide search results?

What kind of search results are you starting from?
This page lists all of the different types of search results that BiblioSpec knows how to handle:
https://skyline.ms/wiki/home/software/BiblioSpec/page.view?name=BlibBuild
 
gwilson9 responded:  2018-02-14 16:07
These IDs come from internal OMSSA software. It does not produce a library file that Skyline will accept but all of the information is available from PSM files. I am pursuing quant vie Skyline because the the number of variable mods makes for quite a long MaxQuant run.

I have looked into your point about scan numbers not matching between IDs and library spectra. It appears that the mgf files I produced are off - looking into this now.