You can use Morpheus to check peptide/protein numbers BEFORE the run is finished!

Maybe you knew you could do this. I didn't.

If you are sitting there thinking "hey...something looks a little off on this replicate.." and wondering whether you should stop the system and run a QC --

---to the rescue!!

I have Morpheus installed on every proteomics acquisition PC in our lab. If you click on the top picture you'll see that I have it set so it can't use more than 2 CPU threads on the PC. It works great for the PC's that have 8 cores -- on the LTQ I only allow it one thread (4 core PC). I don't know that using all cores will push the PC too hard and will crash Xcalibur -- but I'd rather be cautious!

What a revelation,  though!! 60 minutes into the run -- yes -- something is a little fishy here at the MS2 level. Time to stop the run and do something about it so it doesn't impact tonight's samples!

If you don't have Morpheus, you should. You can get it from Github here. And this is the original paper from the Coon lab describing it. 

A pull-down top-down KRAS isoform assay!!

KRAS is a little protein that is a big deal. As one of the most commonly mutated genes in cancer -- and one of the very worst -- loads of people (including a lot of friends of mine) are working on assays to figure out things like:

How much is there?
What mutant variants are there?
What is the ratio of normal to mutant?

With a goal being -- rapid -- sensitive -- and hopefully clinically adaptable!

Most of these assays are digestion shotgun based. Should/could we flip the paradigm and do TOOOOP  DOOOOWN PROOOOOTEEEEEOOOOMIIIICS!?!? Picture any author of this great new paper yelling that out the window of a car as they drive by you. It's much more fun that way.

There are some very good reasons for looking at these proteoforms from the intact protein level. Seriously. Again, it is a small protein. We're probably looking for a single amino acid substitution in a small protein. When we digest that protein and use standard global approaches identification is complicated by a couple of things. The wild type form is often still present AND there can be large differences in abundance between the WT and variant(s) forms.

From the intact protein level we can get 1) the intact mass shift (super important here) and verification from the MS/MS mass shifts. This is 2 points of evidence of the variant -- compared to one single peptide.

Having spent some time looking at this in global data I can tell you that the data processing shortcuts we use in virtually all proteomics software are NOT friendly to these proteins. Protein grouping and
strict parsimony. Remember those? These are trying to make your data report as simple and accurate as possible by making assumptions. Have you ever fed 2 virtually identical FASTA isoforms in and looked at data where both protein/peptide isoforms are present?

You'll get a surprising read-out; if the variant peptide is identified with confidence in run #1 it will appear that only the variant form is detected.

In run #2 if that peptide scores, for example, medium confidence (below your filter cutoff -- remember these are looow abundance proteins) the next run scores that only the WT is present. Protein grouping, strict parsimony, and the use of razor peptides confound your results.

P.S. Quick reminder: If you are still using Proteome Discoverer 1.4 or earlier -- please keep in mind that if you have equivalent information for 2 possible FASTA entries -- PD will always give you the smallest entry (because it is the highest % protein coverage -- this is why you see so many more shortened isoforms of proteins in global data than people using other software. In PD 2.0 or later, the longest peptide gets the nod.

You can avoid all this with top-down!

This team uses a pan-RAS IP (pulls down KRAS, NRAS, HRAS....

....um...thought I'd pull some lyrics from this album to extend the joke....nope. That parental advisory sticker is there for very good reason....yikes....moving on!)

On top of the peptide sequence variation, the terminus of KRAS can be modified in a number of ways --non-mass spec friendly ways.

Now I assume I've convinced you that we'd be better of studying KRAS isoforms/proteoforms with top-down rather than bottom up. But -- will this work in real samples?

This team optimizes that assay for cell lines in culture -- and then gets characterized material from CPTAC -- and shows it works there! I'm a little unclear on the amount of material it took from the fixed CPTAC tissue to characterize it, but I'm assuming we're looking at the normal amount -- what will fit on a slide and show they can determine -- on a precision medicine lab budget friendly Q Exactive, the KRAS isoforms present. As the title suggests -- this wasn't entirely their aim, I guess, they also reveal more basic info about KRAS mutant biology, but that is beyond me. I'm just psyched to have a simple straight-forward KRAS top down isoform characterization assay all ready to go!

The MASTER of time -- Faster Percolator to the rescue!!

Well...I wasted some time this morning because I misread the title of this great new study!!

My first thought was "Master of Time!!! Gotta find a picture of Roger Delgado (he died in a car accident in his second season of Doctor Who, but he and Anthony Ainley look so similar that people who grew up in places where there were more than 2 TV stations (and one that...strangely... just played Doctor Who reruns all the time? Probably think it was the same guy all those years). )

HOWEVER -- FASTER PERCOLATOR!??! Sign me up!

These authors substitute the primary machine learning algorithm Percolator is based on (I2-SVM-MFM -- which is easy to remember with a somewhat childish mnemonic device I won't share here) for a newer algorithm called TRON. They say it works and it works faster, and that's enough for me!

If you don't want to read the paper and just want FASTER PERCOLATOR -- you can get it from Github here!

In depth analysis of the effect of phosphorylation and pH(!?!?) on ubiquitin!

You know what I was just thinking? The ubiquitin process in mammalian cells just wasn't anywhere near complex enough....

...you know what we need? A modification on the ubiquitin modification itself that has far reaching biological impacts on whatever ubiquitination is doing in the first place!

This great new study didn't discover Serine-65 phosphorylation, but they go far out of their way to figure out what the heck it does -- and -- it's complicated.

You won't find much mass spec here. I think they only used it to confirm they'd purified phosphorylated ubiquitin from the unmodified variant. Most of the work here is NMR (i.e., I'm not real confident in any interpretation I could draw from their results, so I won't say too much about them) but pH is also involved here -- yikes -- 'cause it affects the charge on the phosphate group (making me just a little concerned that this isn't something we think about in terms of phosphoproteomics when we chemically enrich them and drop them in formic acid -- are we losing tons of biologically relevant information -- and could our technology ever possibly get us to it -- maybe neutral or negative phosphoproteomics?!?)   And the longest sentence ever award goes to....

This makes sense, though, right? And maybe it is something to think about. Because not only does the phosphorylation change the 3D structures here, but so does the charge on the phosphorylations. Great paper and something I'll be thinking about all day!

Chemistry of easter egg dyes.

Check out this cool article from C&EN about the chemistry of easter egg dyeing!!

Deleted April Fool's Post about how I'm a SWATH believer now.

If you didn't catch it, you didn't miss much, I felt pretty dumb about it a little later. This is much more fun!

Proteomics of gushing wine!

There are times when carbonated wine like Champagne is supposed to look like the picture above.  Example:

However, if you haven't just won an NBA championship or the World Cup or are the heel in a dumb Hollywood comedy, the wine isn't supposed to just go shooting out of the bottle. You're supposed to drink it. When you intend to drink it, but it all goes shooting out of the bottle this is called "gushing" and it is a problem with the wine making process and people want to know what causes it and how to stop it.

PROTEOMICS TO THE RESCUE!! (Abstract link here.)

This group takes a look at gushing wine and the grapes they came from and compare them to non-effected wine/grapes. The work is mostly done with HPLC separation, SDS-PAGE (for relative quantification) and MALDI-TOF identification.

Turns out a sometimes handy (under the right conditions -- tasty) fungus that infects grapes is to blame here and they work their way down to a really convincing protein band that appears exclusively in the gushing wines.

BOOM! Biomarker identified! If you find this protein at high abundance in your grapes or juice you might not want to spend months/years making wine from it.

I like this paper because 1) it's a problem I didn't know existed 2) this team seems to go right out with some of our technology and solve it 3) it's a really clever use of our technology's in kind of a unconventional way. And they walk away with a biomarker!?!?

More evidence, finally -- is it time to question the dogma of the nanoLC?

If you want to identify a protein in a gel spot on an LCQ Deca or QTrap -- you need nanoLC. You just don't get enough signal from your 100 bar Accela LC running 200 (+/-40) uL per minute into that big round ESI source that has the 11 things you have to attach to it.  If you want to identify that gel spot you need to split your flow down to nano and break out the alligator clips and get to work. (If you don't know what I'm talking about, believe me that you are fortunate. There was a time not that long ago where 1 scan per second at 800 resolution was a big deal.)

Okay -- where are we today? 40+ scans per second is possible with 8,000 resolution! We have benchtop FTICR systems that are more sensitive than triple quads when operated with sensitivity in mind.

DO WE STILL NEED NANOLC?!?  It is 2018. Can the mass spec and good chromatography at realistic flow rates be a success for proteomics?  Maybe I'm just optimistic because I've hated nanoLC from the first time I ever heard of it, but I think it really is time. For further evidence:

Check this out -- these authors use a great term in this awesome new paper in press at ACS -- is the NanoLC "dogma?"

This study is AMAZING. Well written, logical, unconventional, and someone in this group is a real chromatographer. This isn't written by a pretend protein chromatographer (like me). They think about confronting this problem from where they should -- from plates to peak dispersion -- working their way through different materials and conditions to get to -- something really really impressive.

The authors didn't reveal their final results in the abstract so out of respect for an awesome piece of work -- I'm not going to give them away here either.  Let's just say that if you are sitting there changing pump seals and dreaming of tossing that nanoLC off the roof of your parking garage -- this might just give you that final bit of motivation to up and do it. There is a reasonable chance that you aren't getting results with that headache that are as good as this team is getting without it.

I feel like I owe the authors of this paper something after reading this. This totally made my day.

PACOM -- A great new Java tool for comparing different proteomics datasets!

Do you have a bunch of proteomics datasets that you need to compare? Or do you have something awesome you've done that you'd like to check against what somebody else put up on a repository like ProteomeXchange?!?

Are these datasets small enough that you can do it within Java memory limits (processed total <4GB, I think? which is plenty of room for lots of processed universal output files)  Would you love a simple step by step interface that gives you instant feedback when you are just pushing buttons instead of reading the instructions?

If you answered yes to all of these the Proteomics Assay COMparator (PACOM) is a great new tool you should check out here.

Seriously -- it's tough to compare proteomics datasets and there isn't much out there in the way of tools that can do it besides Perseus and it is always nice to have additional options. I do recommend you check it out (and read some of the instructions....it looks very intuitive but you can't just sleepily start filling it with mZidentmL(s) without reading anything and keep clicking the forward button and get anywhere...well...maybe you can, but before noon I'm lucky if I can put socks on without breaking something...)

Again -- seriously -- this looks like a GREAT piece of software to have on your desktop for assessing your data, comparing between sets and generating impressive metrics and figures.

Wait -- I got it. WHO NEEDS INSTRUCTIONS?!?! Not this guy! He just needs espresso! Wow. There is some cool stuff in here. Heatmaps, mapping chromosome coverage, multi-set reproducibility metrics(!), word clouds(!?!?) -- okay -- I love this.

PALEOPROTEOMICS Technology Review!

On a scale from 1 to "THERE IS A MOTHER FLIPPIN' T-REX ON THE COVER OF JPR AND YOU DON'T KNOW WHY FOR MORE THAN ONE DAY", how busy have you been the last week or so, Ben?

THERE HAS BEEN A T-REX ON THE COVER OF JPR FOR AT LEAST A WEEK!!! Because....

We desperately needed an update on today's technology on Paleoproteomics, and there is a great review inside. You can check it out here.