Epigenetic variability confounds transcriptomics BUT NOT PROTEOMICS, YO!

Google Scholar alert, FTW!

I recently learned about co-regulation/ co-expression analysis -- this is a tool that genomics people use when they don't have nice annotated pathways to go to. In a nutshell, it's the idea that across a bunch of different situations the proteins/genes/transcripts that are all doing similar things are probably linked together in some way. Sounds simple, right? (Especially if you are well informed and have been doing it for like 10 years before I ever heard of it - good for you!)

In case I get all rambly (this is the paper and it is killer!!)

Okay -- so the central dogma leaves out a bunch of stuff. DNA -> (a buttload of stuff, like epigenetics and weirder stuff like transflips) --> RNA --> (tons of stuff we aren't even close to having a handle on like 150 weird nucleotide modifications and end capping and splicing --> Proteins --> (75% of what we can detect with a mass spec that we have no flipping clue what the flip it is) --> Proteoforms

But this paper is about the epigenetic stuff and co-regulation. Turns out that if you try to make sense of the RNA abundance levels or the genetic localization (where the genes are coming from on the chromosomes) and line that up in some way with the epigenetic variability

-- I mean, it doesn't quite line up. However -- if you download proteomic data it isn't nearly as confounded.

This study uses a bunch of fancy R tools and some really nice mouse data from EnCode (rumor is that they know how to do genomics stuff) and the SILAC data from the SILAC mouse study (from a lab I hear is somewhat reputable) and the results look pretty clear cut.

Transcriptomics is still super useful -- don't get me wrong -- for the price of that mass spec (or 20...) you get an awesome instrument that uses a minimum of $1,200 in reagent per sample and will generate at least 220 GigaBytes of noise and as much as 0.1% of it is even useful data!!  -- but when it comes to correlating with epigenetics, it appears that proteomics does a better job.

**Yeah, I made a lot of jokes. Long day -- but please don't think for a second that this isn't a great paper -- we've got a paradigm to shift -- and this study is an important piece of the mounting evidence that proteomics is critical to understanding biology and shouldn't be sitting in distant second place to those guys with the 100x more expensive/sample sequencing instruments**