The results of this paper at BiorXiV are both head scratching and powerfully thought provoking. After my recent mistake(s) I'm going to try hard to get the results right this time!
To begin -- this is the paper I'm talking about.
Metaproteomics is the use of our normal proteomics tools that can give us a snapshot of information about an entire population of different organisms. I think we're still only just scratching the surface of the potential of this technique and every paper I've seen I think I've liked better than the last one.
In this study, these researchers work with an extremely well characterized mouse model. This model develops colitis. This is an unhappy situation for the mice, but it has been a really effective model for understanding how to better treat the disease in humans. I know some humans with this and it sounds terrible, so I'm on board with the use of this model.
I don't understand the model completely, but this is as close as I'll get. They use Rag1-/- immunocompromised mice and add T-cells to these intestines of these mice. These T-cells are wonky somehow and it causes the colitis condition. A quick GoogleScholar search finds tons of references.
The important part (to me at least) -- we have wild-type mice, Rag1-/- immunocompromised mice, and immunocompromised mice that have ulcerative colitis.
TIME TO DO FECAL METAPROTEOMICS!!
The colon flora of these mice is filtered and the microbes captured -- then they lyse them all and just do proteomics on them (this is all described well in the picture at the top).
A unique aspect of this study over other ones I've read is the use of MudPIT for the LC-MS. They utilize 11 steps using a standard high-low top8 method on an Orbitrap XL system.
Now you've got a RAW file with MS/MS spectra from peptides from a ton of organisms. There is a growing number of tools for metaproteomics, but none may be more powerful right now than ComPIL, which I rambled about here a few months ago. Essentially, though, the approaches are the same -- you are looking for peptides that can tell you what species/genus/or family is present from these MS/MS spectra.
Making the link between PSMs and the genus/family of prokaryotes that are around requires some bioinformagic that I'm not in a rush to try myself. However, I'm pretty familiar with a PCA plot and what that means --
This is a zoom in of the principal component analysis from these 3 groups of mice. Keep in mind -- this is the peptide profile of the bacteria (and other things?) inhabiting the colons of these mice. Mutations in the mice can lead to completely different bacterial populations!
That's weird, right? We've heard a lot about the microbiomes and everything...but... isn't it seriously crazy that changing one gene in a mammal (yeah -- this is an important gene) can completely change the population of the bacteria in that mammal? I'm stunned and I'm not sure what to do with what this implies.
The authors go further than this, of course, they're looking for the bacteria that are linked to the disease they study and they make some interesting findings on the populations and the specific proteins that these populations are really up-regulating. These could turn out to be great biomarkers that could be used to diagnose and treat the disease early.
Seriously -- an awesome study with results I think are both far-reaching and surprising. I highly recommend you check it out.
No comments:
Post a Comment