Tuesday, January 31, 2023



Every couple of years this goes out and I nearly forgot! This takes 90 9 seconds or so (unless you write a personalized letter, which helps) and helps keep Skyline developing and provided as a great unifier for the research community.

Please click here and follow the instructions, cause Skyline is waaaay too annoying to pay for! 

Cyclic ion mobility can resolve closely related phosphopeptides of TAU in real brain samples!


Cyclic ion mobility? That sounds weird, right? I hardly need to use my ion mobility that goes in a straight line and often forget that it is even there.

But what if there are critical phosphorylation sites out there where a phospho on the 8th residue turns the protein on and the 10th and 13th residues all do something completely different?

Maybe it's time to crank up the spin cycle with this thing? 

That's what this group did

This is one of those pieces of hardware I know are out there and it seems like it could have some really cool applications, but I couldn't personally think of one. I could probably use one of my instruments and eventually resolve the three phosphopeptides they did here. Eventually. In a freaking brain lysate? Wait. Does this stupid thing have 4 prolines in it? With enough work you can do anything, but - ouch - now I know I could suggest annoying peptides like this to people like Matt Padula who have invested in this new technology cause here is a real reason to spin it up. 

Monday, January 30, 2023

The need for new biomarkers in ALS!


I've had a scholar alert up for quite a while for ALS proteomics and this nice review puts into perspective where proteomics can have a major impact in at terrifying human disease.

I also have a poster up in my office at Hopkins so it is the first thing I see when I walk in because it is the poster from the long long ago ALS community challenge. I'm going to leave the poster up until this paper is written. When the senior author on the planned paper (and the only person who knew anything about the disease) was lost to it, the study lost all momentum, but the data is really really cool and we had a draft going. 

HEY YOU. Yes, you. Would you like to help finish it? Contact me! I've got tons of notes and an outline and amazing data from like 10 different groups who all found really interesting things in it.

My favorite is a glycopeptide that almost every group identified that, as best I can tell, has never been seen before (from the ASMS Halloween poster): 

Now, there is some disagreement between Byonic, ANN-SoLo, MetaMorpheus, Byonic and MaxQuant as to the sugar monomers (since, you know, the whole sugar monomers having the same stupid mass in a lot of cases and the relative libraries/approaches they used) but the MSFragger team found that after batch corrections this thing is certainly differential. A rotation student in our lab helped me pull out the original spectra and we hand checked them. Definitely real (though I can't say the monomers either). Some groups found a second glycan length at the same place. 

Seriously, check out that awesome review to get motivated and if you want to help those authors reach their goals of finding some new proteomic biomarkers for the disease, a huge dataset is sitting right here. I'll send it right over. 

Sunday, January 29, 2023

THE Proteomics Show Road to Chicago is finally back with Ying Ge!


At long last THE Proteomics Show is back! Wherever you get podcasts! 

What better way to kick off 2023 and the final leg of the Road to Chicago than with the one and only Dr Ying Ge

I think my first question was if she had 50 postdocs given the productivity of that group. She doesn't. How are they so incredibly successful? You'll have to listen to find out. 

Glycoproteomics patterns in alzheimer's disease -- deep analysis of 30 human brains!


30 human brains? 

Multi-enrichment to get to the glycoprotein/peptide patterns? Check this out!

They used "healthy" brains, brains from people who had alzheimer's but were asymptomatic, and some from those who had symptoms at point of death?

Would you be surprised by differing patters of glycan distribution? Me either, but I'd have no idea what to do with those data.

What really shines in this paper, besides the obvious biological importance and loads of work that went into it is how they visualize the patterns of their observations.

If you're wondering what to do with glycoproteomics data this is definitely worth a look. Frustrating things about these PTMs are (among many) the fact that your biology might be that one sugar may be over/under represented in the glycan chains observed. How on earth do you display that?? Like they did. 

Saturday, January 28, 2023

BIRCH -- Identify those batch effects and fix them!


Well....that was a rabbit hole that ended at some old scifi horror show about tree monster. An ad that said the "root of all evil" made me think they probably gave it the correct level of seriousness.

HEY! Remember when we did like 2 condition proteomics? That was cool, because we could run our control, then our treated and spectral count some stuff. No fancy stats, just crap data! (J/k) 

Now, however, we have to do the grown up science with lots and lots of samples. BATCHES of samples, you might say, and now we need to start thinking hard about BATCH EFFECTS.

What's that? Well, that's the differences that aren't your phenotype. Those are the ones that are due to the fact you had a lot more dog dander on yourself when you processed the samples on Monday than the next batch of samples on Friday. Or, if you're in an old dungeon where the HVAC swings your internal temp from 45-95F between those dates....those changes.....

And that brings me back to BIRCH

You can get it here

You need the things above! Your non-normalized file. Your normalized file. A stable internet connection (not sure why) and the software at that Github.

It'll try to sort out what effects were caused by that dog dander when it was 95F in your lab and what was actually due to your CRISPR knockout! 

US-HUPO LFQ Battle Royale -- Sunday, March 5th at 4:30pm!

We have TOO MANY WAYS to do label free proteomics quantification today. 

What we should do is hunt down the most die hard specialists of each one together -- with some folding chairs -- and leave Chicago with at least one less than we flew in with! 

I don't actually know if that is what we're doing, but it is a funny thought. 

I have my opinions. I bet everyone who shows up will have theirs. Maybe we still need more than one method for doing LFQ based proteomics. Do we still need 45? Maybe? But I bet a bunch of us in a room with some folding chairs can sort it out! 

As an aside, it just came to my attention that when I use the "line through text" button on this blog it DOES NOT DO THE SAME THING AS MICROSOFT TRACK CHANGES. 

Friday, January 27, 2023

MS-Ana -- Big DDA spectral library searches with really cool statistics!


I got MS Ana in Vienna at when I was lucky enough to speak at the Proteomics Summer School thing about 1 pandemic ago and have mentioned it on the blog a couple of times. 

And the paper is finally out! 

In some regards, it might not matter to you how much great MS Ana is, particularly if you are using the free versions of Proteome Discoverererer. Starting in the 2.5 (I think?) PD viewer, MSPepSearch requires a commercial license from Thermo. 

BOOM -- MS Ana restores your ability to do spectral library searches! 

Even better? MS Ana gives you cool abilities like choosing how your decoy spectral library is generated. And it generated column after column of statistics on your peptide spectral matches that you can use to evaluate your forward and decoy match data! 

(Mirror plots are awesome) 

Also, are you one of those weirdos who hates Proteome Discoverererer? Guess what, weirdo, there is a rocket fast stand alone version! 

You can get both at www.pd-nodes.org or directly at the MS-Ana page here

I do want to send a special shoutout to this team for helping me with this tool and sharing this manuscript while it was in review. I couldn't have got my first ever paper in JPR without it. (I guess I'm like 14th author on something I honestly did less than 1/14ths of). I also couldn't have gotten that paper without the help of a bunch of other people. The acknowledgement section might have been longer than the paper itself if I hadn't just thanked #MassSpecTwitter as a group.  

Thursday, January 26, 2023

Label free single (big) cell proteomics in 11.5 minutes per sample??


This is a preprint, so all the preprint disclaimers apply here, but Simion Kreimer is just about the baddest mass spectrometrist on this planet so I'm cool with wagering just about anything that this will be in press soon. 

Recently Dr. Kreimer (who might have the coolest job title ever) et al., demonstrated a really fast valve switching double loading method for proteomic sample analysis. It recently showed up in ACS here. And if you've been at any conference in the last 2 years you've probably heard the amazing Dr. Van Eyk talk about how important single cell proteomics is to understanding cardiac diseases.

100 single cells per day! LABEL FREE! 

The ability to discern between cell types! 

Yes, cardiac cells are sort of big, but cardiomyocytes are miserable to work with (95% of all protein is 4-10  proteins!) and this group is hitting 1,000 proteins per cell.


Wednesday, January 25, 2023

Prep hundreds (or thousands) of single cell proteomics samples in a single day!


Want to get going in single cell proteomics? Here is even more resources! 

Check out how unbelievably cool this JOVE protocol is from the Slavov lab!  The video protocol for SCOPE2 is freaking amazing. 

No joke. There is like an 11 minute video explained by the scientists at NorthEastern university -- step by step, that would allow you to prep hundreds and hundreds of single cell proteomics samples per day. 

How did I find the amazing NorthEastern video? I was considering doing a JOVE video and then I felt like a dumb person for not looking this up first. 


Single cell proteomics doesn't have to be terrifying. It can be approachable today and there are great resources to get you going. 

Tuesday, January 24, 2023

Deep proteomics on 875!!! different drugs! DeepCoverMOA!


You know...sometimes I see papers with huge titles like this and think....oh no....what are we overselling today....

This is a study that deserves a big big big title. 

This relatively small group of authors did proteomics on 875 drugs. 

Eight hundred and seventy five. 



שמונה מאות שבעים וחמש


otte­hundrede og fem­og­halvfjerds

walóng daán at pitóng pû’t limá

ثمانية مائة وخمسة وسبعون

ochocientos setenta y cinco

How'd they do it? This figure pretty much covers it! 

Okay, so who cares, right? I'm not going to download 11 million spectra and neither is anyone in the Pharmacology department I work in.

To make it useful, they'd have to have a ridiculously easy web interface that allows you to look at any drug you want and how it affects around 9,000 different proteins. 

Welcome to 

Check this out. I know someone who studies/develops compounds similar to Olaparib. I know every drug has like 12 different synonyms, but -- I'll type the first 3 letters into the compounds box and BOOOM!

Not only can I pick a specific protein to see if it is perturbed by Olaparib treatment, but this resource goes way further than that. It generates correlation plots between it and the other drugs in the library based on the protein level effects! 

This is so so so so so so so so so good. I'm just floored by how much thought went into this. There are, of course, similar things for drug + mRNA levels that everyone uses (most of the biggest ones are old microarray data...blech....) but people use those all the time. Protein level??? Nothing I've ever seen has ever come close. 

Monday, January 23, 2023

Save 19 hours per multiomics prep with BAMM!!


Multiomics sample prep is totally doable, but it takes a long time and can have a load of steps!

Some people in Wisconsin who do a lot of multiomics did the math and came up with about 19 hours to do a good multiomics prep. 

They asked Emeril and he said...that's too long for anything, you need....

...it might get worse...maybe...it depends on what you think of Emeril. (I've made this recipe about 20 times over the years. I sub out the dead bird for grilled eggplant and it is super legit.)

This is what they came up with!

Of course, "multiomics" means different things to different people. For genomics people it is often, "I looked at nucleotides in 3 different ways."

Some of the cleanest proteomics samples I've ever gotten in my life were left overs from where people had done lipidomics and metabolomics extraction and the proteins were sort of just left overs. CLEAN PROTEIN remains. So, this makes all sorts of sense to me.

Total prep time? 3 hours! 

Sunday, January 22, 2023

Overlay your proteomics data on a 3D protein structure with SCV!


Wow, did I ever wish I had this before! 

Have you ever tried to just figure out what peptides you found in a 3D protein context? It is totally possible, but it is a miserable experience (for me anway)!

This is the easiest thing ever! 

Put in your peptides, put in your FASTA! Choose the protein you think it is!

Don't feel like reading? Just run it here!

There is a great YouTube video that walks you through it here.

Saturday, January 21, 2023

Time to adjust that textbook! A whole new set of DNA Damage Repair Proteins (in 2022??)!


You know...I'd be more blown away by this if I wasn't grappling every day with the fact that we don't actually really understand how human cellular division works. (Everything we know has been measured at transcriptional rates! And they don't line up very well at all because not every process in cellular division waits till we build up a whole bunch of mRNA. A lot of it is depolymerizing and reusing what is actually there.


It's 20freaking22 or whatever it is! We obviously know every protein that is involved in DNA damage repair! 

This study makes a very good argument that...no...actually...no, we do not and here is a couple more! 

They do some solid proteomics and then go back and do some even more solid molecular biology. Knockout these proteins and try that experiment again? These things are important! 

Imagine you just got back from the printers the 3rd edition of your textbook you sent out for printing in 2019 and some bozos with a mass spectrometer totally screw up Chapter 4?!? 

Friday, January 20, 2023

You guessed it! Nature's Method of the Year is....long read sequencing...?


For those of us who are maybe occasionally still writing 2017 at the bottom of things when we sign them, you might also be heartened to know we aren't the only ones! 

2022's method of the year is what I HOPE that the students in our department totally think is the only way of doing sequencing, because I thought that was all anyone had been doing for at least the last 5 years.

I'm very confused. I'm glad for the long read sequencing people. Good for (what I hope is like literally all of you, why would you still be trying to reassemble little tiny oligostrings back into a complete story when you can have very long strings of oligos...?) Hey, good for genomics and really good for proteomics. It is MUCH easier to find peptides linked to long read sequencing data. 

I feel like a jerk, but last year's was pretty pertinent and up to date. This seems like one they forgot to get to in 2017, but that's okay, I have things I started in 2017 that I haven't finished. Most pressing is this huge tree I cut down, but it fell into another tree and has just been hanging right over my garage for years. I've had a professional out twice and they've said they'd come back, and then didn't. One day Imma wake up to no garage. I hope, because I'm typing this right now in the office I made in said garage and there is another alternative...

Thursday, January 19, 2023

DIA for PTMs? This group tests DIA-NN vs SpectroNaut vs MaxDIA (etc) for DIA phospho!


One of the last remaining arguments for the dDa crowd like me is PTMs! When that domino falls (if possible), we'll see a full paradigm shift.

A lot of time when people talk about PTMs in proteomics they just mean "phosphoproteomics" and a lot of tools let try to look for them.

How do they stack up? You don't have to do it yourself, this team tried them all out! 

Worth noting, after 3 months of waiting for my university to issue a PO, I just downloaded SpectroNaut 17 and this paper used 16. 

Bonus -- you get to see an HF-X vs a TIMSTOFPro for DIA! 

Tuesday, January 17, 2023

ABOUT f'ing Time! Multiplexed single cell proteomics on an ion trap!


I had a pretty rough month and a shortage of both time and enthusiasm. As I was thinking about not getting up at 4am to have a couple hours to do cool science, I saw this paper in my Scholar Alerts

About f'ing time! I honestly thought I was going to have to buy a hybrid to do this one myself. 

Ions pass BY the Orbitrap detector. They do not strike the Orbitrap detector. As such, they are always at a sensitivity disadvantage to virtually anything where ions make contact with the detector. 

Hey! I learned how to make a GIF! 

You can say a lot of bad things about ion traps. Bad mass accuracy. Low resolution. The spectra from newer ion traps look terrible compared to the ones of the past. Et cetera. But they are sensitive! 

And if you are into the MS3 based quantification methods for TMT/iTRAQ quantification, we've known for a long time that the ion trap can be a serious asset. 

This group broke out the CellenOne and did the NanoPots thing, prepped some real human cells and tried it out. The results are predictably impressive. Higher sequence coverage and more overall proteins identified than with the Orbitrap based quan. However, the quantification distinction at a protein level didn't appear to be as good. (Some differential proteins don't appear significant when using the ion trap). At a functional level the results line up better between the different methods, but there are downsides. 

Worth considering, they used a relatively small sample set. I think they did around 100 cells by the ion trap MS3 method. You'd have to wonder if you'd see the same weaknesses at 1,000 cells. 

This is a great study that I was hoping someone would do for a while so we could see what it looked like and it is really awesome to see this group execute it to perfection.

If you've got an old hybrid sitting around and are on the fence about getting rid of it, you should 100% check this out. I don't have data to support this, but I strongly suspect that the older LTQ Orbitraps with the ion trap on the front of the instrument would be amazing tools for this application. Sure, you can't SPS on them (without getting some magic software from some guys in Boston or altering the software) but you can isolate the most abundant high mass fragment ion from the MS2 and that should work well. 

A really cool thing about single cell proteomics has been how it has sort of flipped the paradigm on instrument selections. Virtually all of the data we've seen from Slavov lab has been on the plain old Q Exactive Classic that data has been right there with literally any other data we've seen from instruments that cost 4x as much. This study opens up SCP to an entire generation of instruments. 

Monday, January 16, 2023

The TransProteomic Pipeline still exists!


Once, long long ago there was this open-ish cloud-like data processing pipeline with a name that would, for some reason, make you think of Santa Claus and hair metal. Not the good hair metal, but the stuff that they play on cellphone commercials and in elevators. 

It was called the Trans Proteomic Pipeline and I just learned it still exists thanks to this new paper! 

There are some really helpful tables inside the paper that tell you what data it takes and what it doesn't from the "dizzying" number of mass spec data formats that we have. 

Honestly, it looks pretty comprehensive. Since the TPP is hosted on someone else's computer, you don't necessarily need a ThreadRipper on your side of the internet to do really sophisticated data processing and manipulation. 

If you are interested in checking out this pipeline and don't know where to get started, check out this page of tutorials! http://www.tppms.org/tutorials/

Wednesday, January 11, 2023

Find the super low abundance proteins with MRMHR (PRM with Zenopulsing)!


This isn't self-promotional because I didn't do this study. Dr. Wheeler did this study. My name is on it because I showed her how to use MaxQuant, introduced her to Skyline and...I guess funding in my name helped pay for her to wrap it up the last 6 months or so. But I guess that's how PI stuff works.

Punchline? Sometimes people who take HIV drugs have negative neurological effects like dementia. We don't know why. What Dr. Wheeler found was that if she took purified endoplasmic reticulum from different brain regions and treated them with HIV drugs, some metabolism occurred. Metabolism is supposed to happen in the liver, not the brain, right? 

While class I metabolism (P450 type oxidation) had been previously observed, class II metabolism stuff really had not been, and that was what she was seeing. And that didn't make sense. 

The QE couldn't detect enzymes capable of this process. DDA/DIA/PRM

The TIMSTOFs didn't detect enzymes capable of these processes (PASEF/diaPASEF).

Targeting on a TIMSTOF is a miserable experience. The files are huuuuuuuuuuuuge. 

Enter ZenoTOF and ZenoPulsing enabled PRM, which they call MRMHR. MRMHR produces tiny output files that are super simple to process in Skyline with ridiculously high sensitivity. 

The first author was able to demonstrate that enzymes matching the metabolites observed in the correct brain ER cell types all lined up. I can't remember if this made the paper, but mRNA could back it up as well.

I know this didn't make the paper, but she also fractionated a bunch of peptides from these brain regions and did really comprehensive proteomics using the same instrument. Using the ProteomicRuler we were able to get some ballpark estimates for how rare the enzymes she is seeing are. Current best numbers are these are at less than 300 copies per cell. Which ain't much. Files are all on MASSIVE at MSV000090576. 

Normalize and integrate your organ-on-a-chip proteomics and metabolomics!


Organ-on-a-chip technology can allow a better representation of in vivo conditions than 2D cell culture can. It clearly isn't as easy as 2D cell culture, but it can also be a good bit easier to scale than animal models. I didn't know it until I read this paper, but the US EPA plans to stop animal testing in their research plans in a few years and these technologies will likely play a big role.

If you are jumping into this field and want to fully characterize your drug (or in this case -- CHEMICAL WEAPON) on your simulated organs with both proteomics and metabolomics, what would you do? I'd just do what this group does!

Hey! I know some of these people...including the...senior...corresponding....author... Wait, that can't be the kid I know. Common name. 

What they do is grow human hepatocytes in scaffold bioreactors to get them to grow into 3D cell cultures. It doesn't look like any funny gelatin things are employed here (whew....all I can see when they grow cells in gels is...their gel monomers...). They take some control little organs and some other ones they treat with "VX" which, according to Google is something like a CHEMICAL WEAPON.

Since they're making tons of these little "organs" they appear to keep their digestions and extractions in plate to keep them scalable. Apparently, though, the size of the "organs" are tough to normalize and since you've got these cells all stuck on scaffolding you can't exactly weigh them. You have to basically lyse them where they are. Proteomics normalization? That's okay, we basically know how to do that (though I'd follow their directions here to make sure that buffers necessary for the "organs" don't interfere with the quan). 

Normalizing the small molecule/metabolomics input? I have no idea how I'd to that. You can normalize the TICs, etc., but that is for relatively small variability. You don't have to load too many samples for LFQ proteomics with vastly different loads on accident to discover the limitations of normalizing at the TIC level, right? What this group comes up with is a rapid method to quantify their metabolomics based on something that reminds me of the name of one of these people...


Joint pathway analysis (!!!) for the metabolomics and proteomics was done in Metaboanalyst, cause apparently it does that now