Real MS3 without an ion trap!

No time to look at this -- crazy busy day, but it is an instrumentation paper and if  @Mass Spec Pro likes it -- it should go here.

Brand new paper at ACS -- hey, David's on it (wassuuup!?) -- and some Makarov guy -- 2 more reasons to post it based on abstract alone!

The abstract says they made the HCD cell in a quadrupole Kingdom trap system behave as an ion trap so they could do real MS3!!  What?!?! I know!

I know some people who do great pseudo-MS3, where they run the sample once, with MS/MS pick their cool fragment ions, then set the in-source fragmentation (SID) to a high level and target the interesting fragment ions they found before --BOOM -- pseudo-MS3. Super useful technique, but -- having spent some time helping with this -- you rapidly find that tuning the SID to get the same fragments the HCD isn't always straight-forward.

Real MS3 in a Q Exactive?

...this was funny for like...3 seconds...but no time to change it now!

MSFragger -- A search engine so good that I'll use command line!

First off -- this image is not MSFragger. This is Fragger, which is a free online video game (think high speed Angry Birds but the pigs are wearing ski masks? -- and all the birds, not just the ones that look like bombs, explode...) that I found while image searching and possibly slightly delayed the publishing of this post.

THIS is MSFragger!

What is it?  It is a ridiculously fast search engine (it says that in the title!)

How fast is UltraFast?  I gave it a 1GB RAW file I converted to mZmL (please pardon capitalization) and it was finished searching it before I had time to wonder if the engine was working. I'm not kidding. It was just done.

If you are like me, sometimes you don't have time to read instructions. Maybe you also happen to like your crooked shelves just the way they are. If you fall in that category, you can get the software directly here! -- But you might want to learn from my completely uninformed mistakes.

First off -- MSFragger runs in the command line. I swear it is worth it, though!  And it isn't one of those command line programs where you have to type in the exact mass of all 6 of your PTMs and if you make one typo the whole thing fails -- it is much better than that.

Once you download and unzip the folder you'll find a "params" file that you can open in Word (Word won't like it, but just say "open any file". Then you have all your settings and mods and stuff. You can set those however you want.  I'd like to mention -- the default mass tolerance range for the MS1 is...500Da!! Yeah. That's how fast this crazy thing is. It delta mass searches by default!

Now -- if you do it my way -- and you've contacted the authors to get help with running the program in such a way that makes it blatantly obvious that you didn't even look at the paper or the nice included instruction manual -- and are now appropriately embarrassed enough to read the very nice instruction manual (I got excited...) you'll need to follow these steps (this is Windows)

Open your CMD prompt (I always type "CMD" wait for it to show the icon then "run as administrator" just so the Windows remembers who's in charge around here)

Guide your CMD over to the folder where MSFragger is.

Cause this was just a first test for me, I put the mzML file I was testing into this folder, as well as my FASTAs.

3 things you'll need to keep in mind before you type that line and get blown away by how fast this thing is --
1) You can't run RAW files, they must be converted to a universal format (I made my RAW into MzmL in Proteome Discoverer 2.2 with no filters)
2) You should use a concatenated .FASTA with decoys
3) You'll need something to look at the PepXML output with

I broke out the Trans Proteomic Pipeline for 2 and 3. This is kind of funny:
1) It took longer to convert the RAW to mZmL
and
2) It took longer to make a concatenated .FASTA file
and
3) It took way longer for me to remember how to use the TPP
Than it took to search a 1GB LC-MS run.

And...somehow I got a lot more peptide and protein IDs than I did with this same file in Sequest! One button press in the TPP and I've got an Excel file to interrogate.

I tried to remove the names of files here, because this is someone else's unpublished stuff.  Also worth noting -- if you use the default params folder for settings for resources to allocate and you've only got 8 cores and 16GB of RAM and you do some serious searches, you might not have enough memory/processing free for your mouse to work. You can easily adjust the resource usage in the params text file.

Tumor xenograft RNA-Seq, proteomics & phosphoproteomics integrated into a solid story!

I don't know enough about cancer biology to really dig into the results of this great new paper...

...but I know a goldmine of valuable cancer information when I see one!   

This study is big. Like 40 tumors grown in mice --> excised, and then put through RNA-Seq, iTRAQ4-plex proteomics and label free phosphoproteomics. It is a lot of data --more importantly,  it is a lot of publicly available data!

The best part for me is the level of integration from the different analyses. This works out to be a cohesive story with phosphorylations that are present making all sorts of sense in terms of the proteins that are regulated by that phosphorylation, or the transcripts that are coming up before the proteins can be translated.

All the proteomics was done on a Q Exactive Classic. Data was processed mostly in MaxQuant, though Spectrum Mill was also used to integrate the proteomics and phosphoproteomics. 

This study was obviously a load of work, but it was executed really really well. Don't believe me? If your institutional firewalls will allow connection with Aspera, you can get all of this data here!

NIH Proteomics Symposium April 26th!

If you happen to be in the general proximity of Washington DC on April 26th and want to see a bunch of cool talks, you can register to attend at this link.

Some people are flying in from some far-off locales to speak and I think it's gonna be a great day. I know I want to see all of these!

SCoPE-MS -- We can finally do single cell proteomics!!!

I seriously can't wait for ASMS! Yes, I'm normally too enthusiastic about the awesome stuff you guys are doing, but -- wow -- I'm just blown away with the jumps you guys are making right now. MetaMorpheus, Prot-Vista, PTM-Oracle are all game changers in my mind -- and I'm gonna stack SCoPE-MS right on top of this list.

SCoPE-MS stands for Single Cell ProtEomics by Mass Spectrometry and you can read about it in this open access paper here.

The paper hasn't gotten through Peer-Review yet, but it will. This is one of those simple and elegant solutions that will immediately make you think "oh...why the Albert Heck didn't I think of that...?!?!"

Boring/possibly incorrect biological significance stuff: Single cell -omics is a big deal right now. When you are mixing a population of cells, lysing them and reading the signals you are getting averages of the cells there. And we often don't want that -- especially in eukaryotic systems. I have friends at different institutions who use completely different approaches to only study certain cell types within tumors (tumors get crazy complex as they differentiate) and if you lyse cells from them indiscriminately you may lose what is going on. Even if you collect cells by laser microdissection that stain the same or have the same surface markers so you can collect them by flow cytometry or something you are still averaging signal. Having one cell -- that's serious...

...Holy Grail stuff!!

Okay -- but the DNA/RNA people have the polymerase chain reaction (PCR) so they are way way more sensitive than us. They really can get one cell, get the DNA/RNA, amplify it and then sequence it. Even with a Lumos, the limit I think I've seen is a decent proteome with 100 yeast cells, right!?!?(can't find the reference right now)

Back to SCoPE-MS:  This technique effectively multiplies your sensitivity by 200 fold!!!

How?? By taking what is generally considered a weakness of isobaric tagging techniques (like TMT/iTRAQ) and turning it into an advantage!

Let's review the TMT workflow first (TMT is a trademark of Proteome Sciences, btw):

1) We extract peptides from out different conditions
2) We tag them with the TMT reagent -- adding the exact same mass tag to every peptide
3) We combine all the tagged peptides
4) We do normal proteomics
5) In the simplest examples (Q Exactive, for example) we get one MS/MS spectra that can be used for identifying the peptide and we get the tag fragments in the low mass range.

For all intensive purposes, the dynamic range of these tags is somewhere between 2 and 3 orders of magnitude. Most processing software will say that if the ratio between your 2 tags differs by over 100-fold, it will give you the maximum ratio of 100-fold. That is default, and I think there have been cases of studies showing linearity much higher than this. Don't quote me on this (don't quote me on anything!), but I don't think anyone has shown linearity beyond 3 logs, even on targeted TMT. Honestly, however, I doubt anyone has looked into it too far -- if your protein is up-regulated 87-fold or 821.4-fold you probably are going to investigate it!

Wow, I have been working on this too long. Y'all -- it probably looks like I throw these together in like 5 minutes (okay, sometimes I do), I gotta wrap this up...someone looks ready to go kayaking!

In SCoPE-MS these authors divert the "normal" TMT proteomics workflow. Only one channel has a proteome with a realistically detectable amount of cells for obtaining a good proteome ~200 cells. This channel is used for peptide identification. The other channels can have as little as one individual cell!  (Please see figure at the very top of this rambly post which explains this better).

You have 10 TMT channels:

One channel is blank for determining noise levels (you'll need it here)
One channel has peptides extracted from 200 cells (a normal...though still impressively small! amount of cells)
The other 8 channels can ALL be from individual single cells!!

The channel with the 200 cells provides enough material for peptide selection for fragmentation and peptide identifications!

Then you can see in the TMT tagging range the relative quantities of the peptides/proteins in the individual cells. You are below the current dynamic range of peptide identification with the single cell channels, but you are well within the dynamic range of the TMT tags!!

How'd they do?

This is the correlation between proteins they found and the results of RT-PCR -- often considered the most sensitive biological measurement possible. Yeah. That good.

They use an Orbitrap Elite for the TMT LC-MS/MS. They do some seriously thorough optimization and downstream statistics. Data processing appears to be all MaxQuant.

Single cell proteomics is finally a real thing....and it is multiplex-able and quantifiable. What an amazing start to this beautiful day!

EDITS: Dr. Norris was surprised by the apparent level of correlation here between the proteome and RT-PCR. This looks substantially higher than what we've come to expect from most studies in the literature. Fascinating thought here -- what if the variability we've come to expect between RNA and Proteome is because RNA levels are more stable among cell averages than the protein levels from cell to cell? If that is the case, it seriously builds the case for how essential this technique is and will be in the future!

Also replaced some incorrect "your/you're/yer".

PTMOracle -- Visualize PTMs(!!!) in Protein-Protein Interaction data!!

I don't have proof yet that this is as good as I think it might be. If it is, this is a serious missing link in proteomics filled in (courtesy of free software!!!)

What I might be ridiculously excited about is the PTMOracle, which you can read about in this new JPR paper here.

If you start doing any kind of PTM study and want to move to downstream analysis you will quickly find you need to:
1) Hijack some genomics tools built years ago for microarrays
2) Use some genomics tools someone hijacked years ago for microarrays and made kinda good at proteomics (and available)
3) Do lots and lots of manual work

Those genomics tools are awesome. They really are. On the downstream analysis/data processing end, they've got a 10 year head start on us.

But -- upregulation of the entire ERK protein may do something very different in a cell than ERK phosphorylation on the normal active tyrosine (or the less common, biologically, but easier to detect with MS - active threonine site) and unmodified genomics tools are either gonna: a) ignore this 'cause it doesn't matter at the transcript level or b) lump these three effects together

Not to say this is bad (and please never trust the accuracy of biological examples I only partly remember and use for examples), but when our instruments have the capability to tell the difference between protein upregulation versus phosphorylation on one or both sites -- simplifying this down to one effect may limit or even mislead the downstream interpretation.

PTMOracle wants to bridge this gap (and, holy cow, the further I get into this paper the more exclamation points I'm adding to the title of this post, I may need to let this cool and copy edit a little.) By separately visualizing the PTMs with this tool we may be able to flesh out these differences and see the real biological patterns. The ultimate goal is to visualize, for example, this treatment shows us phosphorylation of these proteins that protein-protein interaction (PPI) studies have shown us interact -- there is a serious pattern here.

Maybe even cooler? Different PTMs!  I picked the screenshot at the top over the one the authors uses for the abstract (sorry authors and JPR, if this annoys you please email me; orsburn@vt.edu I will switch it) because what if your treatment of cells leads through a pattern of acetylation? What tool on earth visualizes that for you? None I've seen. And, increasingly, these deep proteomics datasets are capable of finding even low abundance PTMs in un-enriched datasets, if you look for them!  You can find that you have an acetylation or methylation signaling cascade that may lead you to the biology.)

How'd they do this sorcery?

They take huge and awesome PTM databases and they take the data from protein-protein interaction (PPI) studies and combine them somehow. There are lots of bioinformagic details, including required formatting stuff. Most of the PPI seems to derive from yeast-2-hybrid assays, but I bet you can only make this better with data from the currently on-going PPI studies (like BioPlex!)

What is important:

1) The tool is live (and downloadable right now here!)
2) The tool is free  (http://apps.cytoscape.org/apps/ptmoracle)
3) They verified that the tool totally works with both publicly available yeast data and with human kinase(!!!) data.
4) If it is 10% as good as it looks in the paper, it is still worth getting excited about. If this isn't the tool that I've been dreaming of since I did my first phospho enrichment, it is at the very very least a step in the right direction!

What is a significant chance in protein abundance?

I'll be honest, this one hurts my head more than a little, but I'm trying to get in front of it for a number of reasons. As I'm sure you're aware, very few people seem interested in just getting a list of peptide IDs anymore. Everyone wants a list of identified proteins with quantification values -- and probably quantification on individual modified peptides. None of our jobs are going to get easier as this trend develops.

It gets worse. As soon as you start talking about quantification -- people are gonna start asking weird stuff about adjusting Greek carrots with pea-values.

If, like myself, this was the first thing that you thought they said. Maybe it's time for us to read some recent papers like this one! 

This paper might seem less intimidating to me than some of the others primarily because it is short. There is plenty of Greek letters and formulas in it. It primarily deals with iTRAQ, but it introduces a lot of useful stats terms people have been throwing around like LIMMA.

From the label free quan perspective (XIC-based, rather than spectral counting -- a whole 'nother topic) we have to go just a little bit further back in time to get some Open Access perspective on what everyone is talking about. There are others, but this is the first reference I'd send someone who asked about this topic.

A big reason for this is that this paper details how the heck to deal with fractionated samples in label free quan. This is not without challenges because there is quite a bit of overlap where a peptide will show up maybe on the edge of fraction 5 and 6 and be present in both. Not to mention a protein might easily have a peptide present in every single fraction. This is the earliest paper I'm aware of that deals with this head-on.  It isn't trivial....

One more -- because we seriously can't even talk about statistics in any sense these days without talking about R packages. Statistics is what R does best, and if you are willing to think something like "hey, I bet the nice people who wrote this cool free stuff knew what they were talking about," this paper is the first step of what a lot of the more recent proteomics R quan stuff is built upon.

Honestly, the paper is a little low on content -- it introduces the concept and this extremely useful website! (http://msstats.org/)

At this site you'll find a breakdown of different statistical concepts, training data sets and instruction manuals -- mostly for the software -- but that will explain a lot of the concepts and why they're important.

Okay, so honestly I don't know if any of this is useful at all. Honestly, I just type stuff. What I do know is that proteomics quan statistics is something we're all going to be hearing more and more about as this field develops and as the commercial software packages more ubiquitously integrate these things -- and these papers/links are things I've found useful.

In case you are concerned that this blog and I are becoming far too serious, here is an angry Bugg puppy dressed as a flying monkey....

...I actually have a good picture of exactly when he bit me for this....

Make quality control more accessible!

How happy is that picture?!?

I stole it from this paper with this awesome title: 

The Proteomics Standards Initiative (PSI) has worked on some great projects over the years to improve our ability to communicate our results (like controlled vocabulary)  and to share our data with conserved formats.

This paper introduces the next project -- Quality Control!!

And it is pretty much just an introduction, this topic is daunting. Do you apply the same QC standards to label-free quan data as you do to an iTRAQ study? If our QC requires new spiked in standards, how do we compare that data to the thousands of historically available studies?

This isn't going to be a trivial project. One thing they really focus on here is that whatever form our future QC metrics take, we'll need to have some sort of conserved data format for the bioinformaticians of the future.

If you're interested in the milestones for the project, you can find them directly here. I can't wait to see what they come up with -- whatever it is will definitely be good for the future of this super cool field.

Awesome review on host-pathogen interactions!

Want to brighten up your Monday when you know you're gonna be stuck working all day and miss the first beautiful day of spring? This might only work for serious mass spec nerds, but this beautifully written open access review started my week of right!!

Awesome figures? Check!
Wows you with the authors knowledge of microbiology, classical genetics techniques, and cutting edge proteomics? Check!
Explains the powers and limitations of a staggering array of techniques and how a complementary technology is needed to fill in the gaps? Check!
(Side note: If you need justification for why the average number of authors/paper is increasing over time, this review puts it in perspective!)

(Stole this figure from the Winnover -- original article here)

I think my main takeaway point from this paper is that there are a dizzying number of approaches that we have now for studying biological systems with mass spectrometers. When they're really sat down in a comprehensive line it is easy to think that -- wow, we might honestly be at a point where maybe only a few people in the world -- if anyone -- could be a master of all of them. But...there are still gaps that all of these tools together can't fill -- and we might need classical and next-generation molecular tools (including some things I've never heard of!) to really get a comprehensive picture of complete biological systems. (I think there are 40 words in this sentence...)

TL/DR: Just a seriously awesome and visually stunning review (prettier as HTML than PDF) that puts our current capabilities and limitations in perspective.

P.S. If you look up the technique in the top image, the first thing you see is this guy....

...which might be kind of hilarious!

Automated digestion on monolithic columns!

Okay...this one seems seriously promising...but it is more than I have time for now.

Putting it here so I remember to read more about it later, but 90 second automated digestions!??! They average 60% coverage of proteins at small as cytochrome C and as big as HSA (~67kDa).

You can check it out here.