I found this reference online:

https://web.expasy.org/peptide_cutter/peptidecutter_enzymes.html

"Pepsin preferentially cleaves at Phe, Tyr, Trp and Leu in position P1 or P1'(Keil, 1992). Negative effects on cleavage are excerted by Arg, Lys and His in position P3 and Arg in position P1. Pro has favourable effects when being located in position P4 and position P3, but unfavourable ones when found in positions P2 to P3'. Cleavage is more specific at pH 1.3. Then pepsin preferentially cleaves at Phe and Leu in position P1 with negligible cleavage for all other amino acids in this position. This specificity is lost at pH >= 2."

I can't remember where I got these definitions, and unfortunately, I did not comment the source in the code, but I am pretty sure they came from either the Trans Proteomic Pipline (TPP) or X! Tandem, or possibly a combination of both.

The protease in Skyline is listed as "PepsinA", so perhaps the A means something and the second-to-last sentence above suggests a case where only FL cleavage is really expected.

Skyline does, however, offer a lot of flexibility in defining your own protease specificity. The definitions in Skyline should be considered examples and good defaults. If you choose <Edit list...> you can see how they are defined and add your own.

A good many peptides "predicted by Skyline", or more accurately, hypothetically possible based on the cleavage specificity of the protease used, are not detectable in a mass spectometer for any number of reasons. Do you have prior knowledge of this peptide, e.g. detection in DDA, a synthetic version you have detected in a mass spec? If not, then you are in a very early stage indeed and should not be surprised to find you cannot detect many peptides generated by insilico digestion and fragmentation prediction.

Thanks for posting to the Skyline support board. Hope this helps.

--Brendan

I think I found the original source in the TPP which appears to have been written by Andy Keller in November 2002 and updated with some new protease definitions (including PepsinA) in 2007.

https://sourceforge.net/p/sashimi/code/HEAD/tree/trunk/trans_proteomic_pipeline/src/Common/Enzyme/ProteolyticEnzyme/ProteolyticEnzymeFactory/ProteolyticEnzymeFactory.cpp

line 73: , {"pepsina", "true", NULL, "FL|-|false|"} // added 2007-02-07, WCH

At least it agrees perfectly with the definition found in Skyline.

If it is not useful to you, however, Skyline allows you to define your own to better match your experimental conditions.

Thermofisher sells a pepsin "smart" digest resin and supports a pepsin cleavage specificity in thermo's biopharmafinder defined as follows

both

C-terminal to

CDEFLMTWY

and N-terminal to

FIMVWY

This reagent works well and is very useful if you are interested in characterizing/quantifying modification at lysines (acetylations, methylations in histones for ex)

(This is a magnetic bead coupled protease, and like its trypsin smart digest counterpart, it gives unrivaled quantitative reproducibility for digestion, and it's absurdly fast (2 x 15 minute digests) and trivial to clean up (remove heat, remove magnetic beads)

(no I don't work for thermo, though I have, but I'm not inclined to them from that experience. These magnetic resins are simply a significant step forward in peptide workflows, the results are better, significantly, soup to nuts, for a marginal increase in price.

All of that is preamble to say, with the pepsin enzyme, you have far more missed cleavages than you might be used to. (You want them, actually, it creates an ensemble of peptides that creates quantitative total coverage (in ion volume) much like shotgun genome sequencing.

For skyline the impact is that for pepsin you need to be able to go to 15 missed sequences not merely 9. It'd be nice to see this get extended. Right now I generate the missed peptic cleavages in a matlab script I built and put them into skyline manually and they pick up the modifications from the peptide settings but because the digest settings are different they are not assimilated into the sequence they come from, so no their protein position is lost.

Thanks as always for building such a useful tool,
Lance