I really wanted to like this new study in JPR. And I might still like it, once I have time to look at the data, and I do really think it's worth thinking about, at the very least, as a concept.
Anytime you're combining global proteomics and phosphoproteomics, you've got my attention. Your idea has GOT to be better than how I do it in Excel for TMT and how I set it up in Proteome Discoverer for label free. Both outputs require a lot of manual investigation and lack those "stats" things everyone is (rightfully) worried about.
In this study, these authors combine global proteomics quantification and phosphoproteomics data. Yeah! That's what we need! And they do the processing in MaxQuant/Perseus with some R. (If the R scripts are included, I didn't see them on a first pass, and this is part of why I'm not loving this paper, I think).
The concept of a "Credible" phosphosite occupancy is pretty cool and it makes you think, right? I mean, you've got some phosphopeptides in that list that are definitely right. And you've got a whole lot that are....meh....phosphopeptides just kind of suck in mass spectrometry. What this group tries to do here is to use some simple algebra to look at the shifts in the phosphoproteome while taking that into account, and it seems to do work. There are pretty plots (actually, all the figures are really nice in this paper) and they seem to support the fact that this is a cool idea, and I'm going to keep thinking about it.
Where I've got mixed feelings here is that I don't think that it is possible to take this same set of data from ProteomeXchange and possibly reproduce what they did here. I've got the supplemental open and, maybe they forgot to upload the pages where they describe what settings they used in MaxQuant? Or Perseus? Or what mysterious R functions? Seriously, though, cool idea, I think.