Monday, January 11, 2016

SILAC analysis provides insight into Ewing Sarcoma

I'll start by being perfectly honest here. There are a lot of terms in this nice new paper that I don't know and don't have a handle on after some light Wikipedia work. Rather than spend my pre-work hours trying to become an expert in Ewing Sarcomas, I'm gonna focus on what I do understand in this paper -- the fact that they did some nice proteomics work!

The paper I'm talking about is from Severine Clavier et al., and is available (open access) here. The sample workup is typical for SILAC and takes place with cells that are apparently an appropriate model for this disease. Separation of the digested SILAC peptides was on a 50cm nanoLC column running a relatively fast gradient and a relatively slow flow rate (~50min effective elution gradient, but at 200nL). My first thought when seeing the 30,000 resolution at the MS1 was that it wouldn't be enough to fully pull out SILAC pairs on a gradient that short, but my first thought appears to be wrong as they report quantification of 1,700 proteins. Not too shabby for a classic Orbitrap!

The quantitative digital proteomic maps (RAW files) they generated were processed with Proteome Discoverer and this is where I get interested. They used Proteome Discoverer 1.3 to get the PSMs and used the 1% FDR. Then the data goes into something I don't think I knew about until this morning -- something called myProMS. (Which you can read about here and directly access here.)

Probably the second most common comment when I give a proteomics talk is "where are the statistics?" And then people get sad when I mention R packages for the stats. myProMS appears to be an R-less statistics package for downstream analysis and, according to this poster, it directly supports output from Mascot, Proteome Discoverer!

How did this study use this downstream software? To generate P-values and volcano plots to pull the statistically significant differential proteins out of their dataset like this!

Does that look sweet, or what?!? To use this software you'll either need Linux or to download a free Linux emulator within your other operating system (instructions on the website)

Hold-on, I'm not done with this paper!  They then take the statistically significant (woohoo!!!) data and then run it through the free STRING network analysis program and find a pathway that makes sense in regards to the morphological phenotypes of these cells and it looks super slick!


  1. Hi Ben. I've been reading your blog for a couple of years now. Great source of information! myProMS was designed to help mass spec people and biologists share and analyze their data together. Version 3.5 will be released in a month or 2. It will handle Proteome Discoverer 2+ and MaxQuant data formats!
    Patrick Poullet

    1. Patrick, thanks for the info!! Please let me know when the new version is released. This looks like great software and something that a lot of people ask for and about.