Hi Alex,
There are multiple ways to achieve what you are after without having to remove your target protein(s) from your background proteome.

For years it has been possible to use Edit > Unique Peptides to determine peptide uniqueness for any protein in your document. Just select the protein and the Unique Peptides form shows you what other proteins in your background proteome share the peptide. In a case like the one you describe for a single protein this should probably be enough. If only your protein of interest shows up in the form, then you have all unique peptides. Otherwise, it is easy to un-check the peptides that appear in other proteins.

Currently in Skyline-daily, Brian Pratt implemented a new feature which (after years of requests for this) makes it easier to do this automatically and on a much larger scale. In Skyline-daily, you will find a new option "Enforce peptide uniqueness by" associated with the "Background proteome" control in the Peptide Settings - Digestion tab. When you have a background proteome, you can then choose from Protein, Gene or Species in this list. It is expected that your target proteins also appear in this background proteome, for the algorithm to determine which Gene or Species they belong to. Otherwise, the algorithm probably just defaults to determining whether anything in the protein appears in the background proteome at all, assuming that everything in it represents different Proteins, Genes and Species.

Good luck with achieving your desired experiment. These options should help, we hope.

--Brendan

Hi Brendan,

I have a question regarding the 2 features you have described above. In the meantime, the option "Enforce peptide uniqueness by" in the Peptide Settings; Digestion tab is fully implemented in Skyline. When I apply this one to filter my peptide list according to a background proteome, my peptide candidates get less and it seems to work nicely. However, if I then check the list protein by protein with the "old way" via Edit > Unique Peptides, there are still (few!) not unique peptides appearing in the columns, even so the same background proteome should be used. I can then kick these few out but it still confuses me... Am I doing something wrong here?

Thanks a lot in advance and greetings from Germany,
Roman

Roman,

Can you send us your Skyline document? I will try to figure out why you are seeing what you are seeing.

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

Hi Nick,

I have just uploaded a zip-file (only a target peptide list, no XICs) to the folder you suggested under the file name 180509_Roman_for_Nick.sky.zip. Somehow my description was not uploaded, at least I cannot see it there now, therefore here a couple of details:

There should be 2 background proteomes, one canonical human proteome and one containing isoforms, both from UniProt. The described issue appears independently from the proteome I use. If I filter my protein FASTA against the canonical background proteome with my filter settings and enforce peptide uniqueness by protein, I end up with 15 suggested peptides (this is what you should see when you unzip the file). However, when I check via Edit-> Unique peptides..., 2 of those 15 candidates seem not to be unique. With the isoform background proteome I am down to 6 candidates, and still one of them appears not to be unique if I double check.

One other question concerning the global uniqueness filter found in peptide settings: if I don't enforce peptide uniqueness at all, I have 19 peptide candidates. However, if I go for "Enforce by Gene", I still get 19, so nothing changes! However, there are peptides there that are shared by proteins that are expressed from different genes. Is this information not properly included in the background proteome or why is the gene criterion not working for me?

Thank you a lot in advance!
Roman

Hi Roman,

As you suspected, the problem is that there is no gene information in the protdb file. May I see the FASTA file you started with so we can understand why this happened? Uniprot server interaction can be a moving target, though I'd have thought we could pull what we need just from the description in the FASTA.

Thanks

Brian Pratt

Hi Brian,

sure, I have just uploaded Uniprot_Human_Canonical_20303.fasta under https://skyline.ms/files.url. The isoform-containing FASTA is way bigger, do you want that one, too? Would it be helpful to have the FASTA from my target protein as well? If there is no gene information, this would not be a big issue for me, I always enforce uniqueness by proteins anyway. However, the discrepancy between peptide settings filtering and Edit-> Unique peptides... filtering still confuses me.

Thank you for looking into this!
Roman

Hi Roman,

That's a new variant of FASTA header for us - we'd not seen the OX=xxxx entry before and it tripped us up before we got to the GN= entry.

We'll make sure that works going forward, but for now if you just open that FASTA file in a text editor and replace every "OX=9606 " with "" you should be fine. (You'll have to remake the protdb file of course).

Thanks for letting us know! Community feedback is how Skyline improves.

Brian

Roman,

The reason that the peptide "GCSVK" is considered unique is that there is only one protein for which it is a fully tryptic peptide.
The peptide appears in other proteins, but it is only semi-tryptic in those other places.

The Unique Peptides dialog shows you all proteins that contain the peptide, regardless of enzyme. The uniqueness filter only considers other proteins that could produce the peptide using the enzyme that you have specified.

What behavior would be most useful to you for your experiment? Do you think that the peptide "GCSVK" uniquely identifies that protein, or would you rather that Skyline consider it not unique since it could be produced by other proteins being cleaved by other mechanisms?

Hello Brian, hello Nick,

@Brian: Ok, thank you for clarifying why the gene uniqueness does not work here! Indeed, OX is the species or taxonomy indicator for UniProt, 9606 is human.

@Nick: Thank you! Now it seems so obvious, since the background proteome is in the Digestion section of the settings! I also agree now, that for us "GCSVK" uniquely identifies our target protein, since it is not fully tryptic in other sequences.

Thanks to both of you and the whole team for developing this great piece of software and making it available free of charge (even though, in my opinion, Skyline outperforms many of its commercial competitor software packages). As members of a proteomics core facility from Heidelberg University, we have introduced Skyline training courses for our local customers last year and they really like it so far.

Roman