Is proteomics seriously about to realize all the promise we've always thought it had? You guys are doing the most amazing stuff right now! Mind blowing, breath taking work is pouring out of the literature faster than I can possibly read it!! I'm still absorbing the fact we can do proteomics on 1 (ONE!) human cell and....this new Nature Protocol study...
...shows a complete workflow to quantitatively map protein expression at the subcellular level!
I feel like my enthusiasm might be making this post a little jumbled. To clarify: hyperLOPIT is a different approach to extremely low level analysis than Scope-MS, but it might be just as powerful.
I'm not going to do this justice in the limited time I have this morning -- but this is the gist -- this procedure allows you to see where the differential regulation is in your proteins -- at the subcellular level. Subcellular fractionation by gradient centrifugation leaves behind specific markers that allow you to say where those proteins originally came from.
Definitely not doing this justice: (Contain enthusiasm and maximize limited time, Ben! Typing to yourself isn't helping...)
Have you ever used gene ontology (or protein ontology via Annotation in PD) based on cellular localization? What you'll find is that a ton of proteins are annotated at: Cytoplasm, membrane, nucleus. This isn't bad data. At a biological level we know that a lot of proteins have functions in multiple cellular compartments. Maybe this proteoform ends up working in the nucleus and this one does stuff in the cytoplasm.
In our normal shotgun work, we bust it all up, mix it up, cut it up so we don't even know what proteoform it came from and then we check for up- down- quantification. SUPER VALUABLE info, don't get me wrong.
But...hyperLOPIT can add a whole new dimension to this, by allowing you to see the differential regulation and WHERE it is occurring in the cell! See how much more info that is and could be?
Does the data processing sound like an impossible nightmare?
These authors provide (can't check it out yet) open source software using advanced machine learning algorithms that will process the data directly from this procedure!!!
One more crazy feature. This procedure uses TMT10plex! They use SPS MS3 quan for absolute certainty in their measurements, but at this level of fractionation and separation I think ratio suppression wouldn't be terrible in MS2.
Proteomics -- HyperDRIVE!
Think this blog is weird? Me too. But somebody out there took time to make a Boston Terrier(?) in hyperspace. Makes me feel better about it!