Isobaric labeling technologies like TMT and iTRAQ are awesome cause they save us a ton of time. The down-side has been traditionally thought to be that: "once you -plex, you are done". You TMT 10plex ten samples and you can compare those, but you can't compare anything else. Several studies have shown you can go beyond that with internal controls and things, but I'd argue that this is definitely the biggest one!
The study is from Ornella Cominetti et al., and is in this month's JPR here. They took plasma from 1050 patients, depleted it and TMT 6plexed it on an Orbitrap Elite (separation was on a 50cm 2um particle columns) and used PD 1.4 as their interface for Mascot and Scaffold for X!Tandem. The downstream work was done with R and GraphPad.
They do a really remarkable job setting up the experiment and sample groups (nicely randomized!) and I'm nothing short of impressed with the downstream stats. Good experimental design, great chromatography and MS method, and thorough data searching and setup = one solid paper!
How'd they do with a set this big? Direct quote from the abstract: "We demonstrate that analyzing a large number of human plasma samples for biomarker discovery with MS using isobaric tagging is feasible, providing robust and consistent biological results"
Number-wise? Its pretty sharp. They quantified hundreds of proteins that had NO missing quantification points across all plasma samples. Over 1,000 patients! In the discussion they detail why, for a cohort this big, they would stick with this method over a data INdependent (DIA) type approach because they'd be stuck with 6x the run time even if they would get a few less missing values. All along, just a nice solid study. One day I'll be less impressed when people show me proteomics from thousands of people, but that day hasn't got here yet!
No comments:
Post a Comment