I'm on my second read through of this new study in JPR. After the first one I realized my coffee was sitting on the counter untouched. Having rectified that, it is starting to make sense. Why would I read it again? Because I wouldn't mind having like 50% more confident phosphopeptides and -- like you, I bet -- I've always had some skepticism when I see at the PSM level multiple phosphosites observed that I don't see in my final peptide/protein report.
The question addressed in the study is in how we're "rolling up" peptide spectral match data when multiple PTM (here, mostly phospho) sites are possible. And ALSO how are we dealing with these same data in regards to FDR.
I just tried and realized that I don't have the time to do this justice, and it is better described in the apper -- and I should be getting ready for my daily death-defying commute through the Thunderdome.
The important thing here is that 1) it doesn't look like we are doing it right (taking just the best scored PSM for each potential PTM site and rolling that up) 2) it doesn't look that hard to implement what this group did and 3) this is very convincing that we're both losing information AND it this is one method that might allow us to get it back. Super cool study and worth at least one readthrough.