One very convenient shortcut in most proteomics workflows is the fact that we basically ignore individual genetic variation.
If you are doing proteomics on perfect clonal populations like E.coli K-12 strain or Mr. Meeseeks
You don't have to worry about missing a peptide spectral match (PSM) because of an amino acid variant (AAV) [let's assume mutations don't spontaneously arise :) ]
For everything else? Missense mutations/alterations leading to single amino acid substitutions from one organism to another are all over the place!
GenPro is new software packaged described in JPR here that can create a personal protein database from whole exome sequencing data! Emphasis here on the second part -- exome sequencing is the much cheaper genome sequencing type. (We just had ours done by a new startup and I think 125x coverage was $400 USD direct-to-consumer.)
First off -- this isn't the only way you can make a database like this, by any means. However --
1) GenPro appears to be the simplest way to do it I've ever seen
The software is available on GitHub here with step-by-step instructions (it isn't a GUI -- we do have to type/copy/paste a lot of stuff).
2) GenPro is the most thoroughly validated one I've seen so far!
They pull BRAIN tissue and do LFQ global proteomics compared to the GenPro generated database and find nearly 1,000 new peptides!
Okay -- if the authors or editors see this and want this image taken down, please email me (orsburn@vt.edu). You'll get a swift removal and heartfelt apology, but I LOVE this figure and I think everyone should subscribe to JPR and read this study enough that I'm going to risk it --
Check this out!!
This is HRAM fragmentation on their (OT Fusion, I think). The top is the normal variant -- the one that you'll find a PSM for using UniProt and from a normal brain of this species of organism. However -- in the personalized database there is an AAV -- and it is very clearly picked up from the MS/MS spectra. Just to cover their bases, these authors produce a synthetic standard and it shows 100% that the variant is there and fragments in that way.
These authors make some interesting biological conclusions from their observations on the AAV distributions, but you'll have to read it to see what they are. They make some heavy peptide standards and go back with PRM to obtain absolute quantification on some of the 1,000 variant peptides they find.
TL/DR?
1) Great new piece of free software
2) Thorough validation of said software
3) Thorough mass spectrometry from some people who obviously know what they're doing,
4) Extra steps taken to strengthen their observations and fill in some new insight into brain biology.
5) Shows how close we are getting to personalized proteomics!
No comments:
Post a Comment