Despite all the phosphoproteomics work that has been done -- working with phosphopeptides still kinda sucks. This new paper in JPR takes a whack at improving one of the pain points at the data processing level.
The strategy is worked into X!Tandem Pipeline 3.4.1 (which is freely downloadable here. Don't worry, the interface isn't in French!)
It employs a new (at least to me!) grouping strategy for both identified phosphopeptides into "phosphoislands". There are many reasons why this is a good idea, but the one that is the most obvious is how we deal with miscleaveages or peptides that show up as (for example) both +2/+3 species. At the proteoform and biological level they are representative of the same thing -- and increase in phosphorylation at this one site on the protein. If you've done much work with data processing of PTMs -- you know this isn't going to be how your output looks in any software I use. It is going to look like separate events and, while you can figure it out, it can be an annoying exercise!
By grouping these events into a single event -- a phosphoisland -- it takes a lot of work out of figuring out you're looking at upregulation of phosphorylation at this one site. Another advantage is when don't have great evidence for the actual phosphorylation site on a peptide where multiple modifications could occur -- it looks like you could group those together as well!
Worth noting -- this isn't the only power this snazzy interface has -- it has easy pull-downs to see your sequence coverage and it is grouping peptide/protein IDs really logically as well!
No comments:
Post a Comment