Tuesday, December 31, 2013
Recently, I've gotten to be in on a couple of Orbitrap Fusion installs and got to play around a little with different samples. I've uploaded some methods that have produced nice results to the Orbitrap methods database. The first is an MS2 only based method for iTRAQ and the second is the same for the TMT 10plex reagent.
In addition, I received some RAW data from a lipid researcher and uploaded his method for looking at lipids on the Q Exactive. As always, treat these as a starting point and email me (firstname.lastname@example.org) if you have questions/comments/suggestions. These are constructed in my free time and not meant to be considered as the end all, but for at least these 2 Fusion runs I'm pretty happy with what I got out of them!
Monday, December 30, 2013
Proteome screening of pleural effusions identifies galectin 1 as a diagnostic biomarker and highlights several prognostic biomarkers for malignant mesothelioma.
Mesothelioma is a cancer that starts in the protective lining around your organs. Since the massive removal of asbestos in North America over the last 20 years or so, we rarely hear about this cancer now, as it was the major cause. Asbestos was pretty useful stuff and since removal is expensive, there are lots of places where this cancer is still a big deal. Unfortunately, there aren't good biomarkers for clinical assays.
Until now! In this study, a team of Swedish researchers in conjunction with a medical institute in Turkey track down a series of good biomarkers for this nasty disease. The experimental method was simple. They used iTRAQ 8-plex to label (top-14 depleted) pleural fluid from Turkish patients with various lung conditions. They used hi-resolution isoelectric focusing (see, the OFFGEL isn't dead!) to pre-fractionate the samples before running them on an Orbitrap Velos.
Note on the method: The method employed is one that could be considered suboptimal. There are several places where the fill times and method could be tweaked to improve the theoretical peptide coverage and quantitative accuracy. For current recommendations on running iTRAQ samples on an Orbitrap Velos, please see the Orbitrap methods database in the right hand of this page.
The results here, however, are extremely nice. Multiple biomarkers were identified and validated with statistical significance which can only lead to an improvement in the correct profiling of the patients afflicted with these diseases.
Sunday, December 29, 2013
Oh, Lists! Why do we love you so much?
In another example, Google+ suggested this article for me regarding the most active Twitterers (I don't think that is the best diction, but you probably know what I mean) in our field. I follow most of these people, but will probably start following more. I ran low on topics to write about over the holidays!
You can find the list here!
Tuesday, December 24, 2013
Monday, December 23, 2013
Use of quantitative mass spectrometric analysis to elucidate the mechanisms of phospho-priming and auto-activation of the checkpoint kinase Rad53 in vivo
Rad53 is a protein that is involved in DNA repair in yeast. An extremely similar protein in humans, Rad51, is shown above. You irradiate some cells and probe with an anti-Rad51 antibody and you get the distinct foci shown in green above.
Despite years of research, DNA break repair is something that is poorly understood. A really nice paper in MCP uses a spiked in SILAC approach and high res mass spec to try to pull back the curtains.
In this multi-lab (heck, multi-nation) study, this team uses Rad53 deficient yeast strains and MMS (methyl methanesulfonate, a chemical that induces DNA double strand damage) to try to fish out the pathway leading to Rad51 activation.
Minor comment on the paper. The MS1 search tolerance was set at 25ppm in the MaxQuant/Andromeda runs. Particularly in a SILAC study, I think that this window is a little too big and might lead to mismatched pairs and maybe a raised FDR. Otherwise, this paper is a nice solid look at using a good classic genetics (knockout) approach coupled with HR-MS to fish out pathway differences. This is also one of those rare systems approaches successfully using spiked in SILAC.
You can pull up the original paper while it's still open access here.
Friday, December 20, 2013
I have no idea what is happening in the image above. GoogleImages gave it to me when I said "pug translator".
Anyway! As long promised I finally put up the first version of the mass spec terminology translator (see top right). This will be evolving. I had a 6 hour flight with no internet yesterday and this was the first list I came up with. As new jargon (and more importantly, jargon that I just haven't thought about yet) pops up I'll continue to add on, and hopefully clean up. A bunch just popped into my head while I was writing this. Expect expansion.
And a direct link here.
Lets start off like this: This is a paper in MCP but they didnt do any mass spectrometry, not even a little, so why do we care?
Next article please!
No, wait! This is awesome, I promise.
You know how I'm always preaching that the database ought to match the size of the thing we're doing MS/MS on? Why on earth would we go to all the work of cutting a liver out of something, doing proteomics on it, and then searching all of the proteins that mice can produce? Does that make sense? Nope! But there isn't much we can do about it.
But what if somebody took the genome and cut it down to what proteins are expressed in what tissues by doing RNASeq on individual tissues?
And what if they backed it up with protein arrays to show that the protein levels strongly correlate with the transript levels?
In my (humble?) opinion, thats a paper that ought to be in MCP!
Who needs a mass spec? Wait! What? Everybody! But this is pretty cool and will be helpful later!! Lets try re-running that brain, liver, or whatever with a FASTA that only contains proteins from that organ and see how our results jump through the roof and our FDR drops through the floor!
Thursday, December 19, 2013
I knew this was coming. I really really did. But that doesn't stop me from being crazy super psyched about it.
You can buy a Byonic node and put it directly into Proteome Discoverer right now!
If you don't know about Byonic, go to the search bar and look up my ravings about this software over the last year. Next generation search engines are GO! MSAmanda for high-high MS/MS data and Byonic for your PTMs and in 2013 Proteome Discoverer has went from relying on (perfectly suitable) search algorithms written when I was in high school to brand new engines that 1) take adavantage of high res MS/MS data and 2) can search PTMS in a completely new and super efficient way!
I don't have this yet. Expect me to hunt it down and data to follow!
Tuesday, December 17, 2013
This article came by way of this month's Wired (which is pretty great, btw, as Bill Gates stepped in as editor). I knew about RocketHub, but I didn't realize that people were successfully using this as a method to fund their research. You can browse through, look at research proposals, and donate to ones that you feel are viable.
I flipped through a few on the site (by simply using the search term "science", which also brought up a lot of science fiction). LabScene is one of my favorites, a project that hopes to replace the peer review process!
You can go directly to Rockethub here, and check out the video for LabScene here.
Monday, December 16, 2013
This is a neat little app. Visually appealing and with nice little summaries of mass spec techniques and hyperlinks to where you can actually purchase the consumables to do the work. Now, how many people have purchasing departments that will allow them to use an Ipad to place orders? I don't know, but we all know everything is moving that way so it's only a matter of time.
Sunday, December 15, 2013
We all want to cross-link peptides and do mass spec on them, right? We have a protein of interest and we want to know what other proteins are interacting with it. So the strategy is to throw in one of the 20 or so crosslinkers out there, pull down our protein of interest with an antibody or something and then do MS/MS on everything OR specifically study the crosslinked peptides.
Problems? When we pull down one protein we pull down tons of proteins. Part of the reason is that antibodies really aren't 100% specific, particularly due to the incredible number of protein isoforms present in biological systems. Another part is that no protein system is composed of just a few proteins. Billions of years of evolution have forced an unbelievable level of intricacy in hundreds if not thousands of proteins working simultaneously together to efficiently achieve even the most simple of tasks in the most energetically favorable manner. This isn't done by textbook pathway drawings of 6 proteins. Not when throwing in 100 more will could the energy requirements of that reaction by 30%.
Another problem? The false discovery rates of small protein complexes (or, heck, even big ones) sucks. FDR works best with bigger and bigger datasets. Small ones just don't work right.
A worse problem? The crosslinked peptides give horrendous FDR calculations. Awful. Cause you have to use so many dynamic modifications per peptide sequence. This equals horrible dynamics. Add that to your small sample size and your often looking at a random number generator.
This paper is badass, btw. You know what they did? They look at the crosslinking in a biologically relevant context. No kidding! They take into account the protein crystal structure providing the proximity of the residues for crosslinking and throw that into the FDR!!!! Cause we have that data out there for most proteins (okay, not most, but for most important proteins.)
Okay, so there is a disconnect here, maybe. Yes. We have the crystal structure for individual proteins. Lots and lots of them. And this process will work for that (they prove it in this paper using RNA Polymerase II as an example). But what I'm more interested in is in complexes, and we don't have nearly the same degree of data for those. So I guess I'm extending the real power of this paper a little, but what a step forward! I'm imagining the extension of this algorithm if it eliminated binding sites we know are in use or are deep in the internal structure of the protein. But, holy cow, this paper is really really smart....
Read it (currently open access) here!
Saturday, December 14, 2013
The image above is stolen directly by/from Google Images. First thing that pops up if you look up "spectral counting". Anyway, I often get questions about using Proteome Discoverer for spectral counting. I have some slides that I cut from an iORBI I attended several years ago showing spectral counting in PD and this is what I send people.
I'm not a big fan of spectral counting, but it does have its place sometimes. The slides show you a comparison of spectral counting vs. quan of peaks at the MS1 intensity level as the limitations in dynamic range from spectral counting (with a very nice reference).
EDIT (2/1/17) New DropBox link!
Friday, December 13, 2013
My first Xmas present arrived while I was on vacation. Peaks 7 came out and I got a nice long trial license to check it out. I installed it on my plane ride home.
The list of new features from Peaks 6 to Peaks 7 is kind of mind blowing. This software is very very sophisticated. It is supposed to be faster, it has online collaboration and sharing modules embedded within it and it now does label free quan (with really pretty heat maps). There are a full list of the new features available here.
However, the features I'm most interested in checking out are described as "improved de novo localization scores" and "statistical charts for accurate filtration of de novo data". I love to hear about improvements in scoring accuracy and FDR, and de novo is where FDR needs the most improvement. Anybody going out of their way to improve that can send me their software in an easy to install way with a nice free trial and I'm going to do a fair job of checking it out in my free time.
Bonus? You can directly import MSF reports from Proteome Discoverer 1.3 and 1.4. I don't know if this will simplify my workflow for de novo studies (link to the video I made for working de novo into PD), but it just might. I'll be back later with first and final impressions!
Thursday, December 12, 2013
This is a cool resource I recently stumbled across! The MaConDa is a really easy and simple site that has exact mass information on previously identified contaminants in MS/MS runs. I know most of us have the supplemental information Excel sheet somewhere from that cool paper from a few years ago lying around somewhere, but this is like that with a couple of neat twists.
1) You can filter by instrument type (ion trap, QQQ, or TOF [probably what you'd use for Orbi])
2) You can set a custom PPM error, such as that for your instrument
3) You can filter by contaminant type
4) You can output a list that contains adducts. Icing on the cake!
Check MaConDa out at this link. I bet you'll end up using it sometime. I used it today!
Sunday, December 8, 2013
Anyway, a new to me tool for gene ontology is GOrilla, which appears to be hosted by the Weizmann institute. I know there are a lot of GO tools out there, but this one has a couple of nice features beyond its great name. The first is a customizable p value threshold for your gene enrichment analysis. The second is the easy control of your output format. You can simply checkmark the box to output your data in Excel format and/or you can export it directly into the Revigo visualization tool.
You can check out GOrilla here.
Friday, December 6, 2013
Ive got one backpack full of stuff, my IPad, and I'm on a bus to the mountains of northern Japan. My goals include seeing a wild snow monkey, snowboarding, and making a dent in the global supply of Sapporo (which is crazy cheap here!). As a consequence, the blog may not see many updates until I return. It should be an exciting winter, however, as many many good things are happening and I can't wait to share them with y'all!
Thursday, December 5, 2013
Are you running some sort of quality control when you do proteomics? If your answer is "of course" then I like you. Heck, I'm a friendly guy. If your answer was "never, and I hate pugs!" I'd probably still like you, but I might like you better if you are running some sort of QC.
We need to have some sort of metric of how our instruments are running. I'm often asked what my favorite is. This question commonly comes after people find out that I don't know what a BSA digest should look like...
The answer, and I swear I wrote all of this a long time ago, is PRTC.
PRTC? Now, I'll admit, I didn't know about this thing until I joined Thermo. But I like it so much that I keep aliquots in the ziplock baggy that I keep my toothpaste in for when I travel.
PRTC stands for "Pugs Rock The Cazbar!"
or Peptide retention time calibration....
What it is: A clean, equimolar, mixture of 15 isotopically labeled peptides for varying hydrophobicity. Running these can give you a picture of the performance of your LC gradient and your signal intensity over time. Since they are isotopically labeled you don't have to worry about them being mistaken for something else if you happen to have a small fraction carry over into your next run.
They are also well aliquoted. So you can use them as spiked in standards at low concentration to normalize label free peptide quan from sample to sample. (I should have some nice data on this in the next couple of weeks.) And in Pinpoint, you can simply add in QC peptides and it throws them in.
Oh. And it's cheap.
Downside? None. Period. Exclamation point.
Want to know more? Check out the product page at Pierce, or this sweet application note written by some of my favorite people!
Wednesday, December 4, 2013
This is a nice analysis that comes from a pretty simple set of experiments that were just done nicely. The article from Mostovenko et. al., (open access!) compares multiple methods of fractionating both an E.coli digest and a single digest of human plasma.
The methods compared are: SDS-PAGE vs. SCX vs. IEF. The output is unique peptides and overlap between methods. Interestingly, in the bacterial digest, SDS-PAGE and SCX run kind of neck and neck, while IEF lags behind. However, in the plasma digest, SCX fractionation is the clear winner.
Tuesday, December 3, 2013
This is a paper for all you bioinformatics people out there. Partially because you need to have a stronger background in computer stuff that I do to even install and use PECA.
It appears to be a nice tool for the analysis of RNA expression data and it has the capabilities for also inputting quantitative proteomics data from different formats.
PECA differentiates itself from other software by specifically targeting transcripts or proteins that fall within a steep range but limited range of almost logarithmic increase or decrease.
For more information, check out the abstract at JPR (not open access) here. You can also download the software for compiling and install directly from sourceforge here.
Monday, December 2, 2013
In press at MCP is a great new paper showing how arginine phosphorylation is used by Bacillus subtilis in the regulation of response to stress.
And it isn't a little involved. It's a lot involved. This study shows that it can be linked to heat shock response, response to oxidative stress and in the resistance to drugs. Pretty impressive findings out of well-characterized model organism.
Beyond the fact that we have yet another PTM to worry about, this paper is valuable for the clear (and pretty simple!) methodology for harvesting, enriching, and analyzing arginine phosphopeptides.
A good read, if only for putting in perspective how much we don't know about the physiology of even the most well-studied organisms! Definitely check it out while it is still open access. Direct access to the PDF is here.
Sunday, December 1, 2013
The human proteome project has been rocking for a while now. How far has all this work gotten so far.
Well, here is an update (not open access), compliments of JPR and Terry Farrah et. al., and the number is around 62%.
Oh, 62% of the coding sequences of DNA that we think code for proteins have strong supporting evidence of their existence in MS/MS spectra. That's pretty cool right Over half way!
Let's take a moment and think about how great this is, and how far we've come so far!
While doing so, let's forget the fact that one post translational modification can have dramatic ramifications on the function of a protein. Let's also forget the fact that in 2011, we knew of about 80,000 specific PTMs. Also, let's forget about conformational changes that can have effects every bit as impressive as PTMs.
Please don't get me wrong, I'm not trying to put down the work of the participants of the human proteome project or the good people at ISB who are running the peptide atlas. I'm simply concerned about our tendency to underestimate the complexity of biological systems. We did that with the human genome project. First of all, getting MS/MS spectra for all of the proteins predicted from the HGP data is the tip of the iceberg. Secondly, let's not declare big ongoing projects completed for a while. Grant dollars are pretty scarce out there, and we don't need ignorant politicans reading headlines and cutting all the money to our friends because they think the job is done.
Ran into this one thanks to Twitterer @PastelBio