NIH BioArt has a Q Exactive!

 

If you're rapidly drafting some figures - Biorender can be super cool. But what if you don't have a license? NIH BioArt is adding stuff all the time! 

https://bioart.niaid.nih.gov/

Proof? There is no way that it had a Q Exactive when it first made my blog last year! I would have found it, but now it has it and more free/open art to use to really drive that dumb point home that you're trying to make! 

While you're there check out how much cool new stuff has been put on NIH 3D! 

https://3d.nih.gov/

Mandatory reminder that these government resources are probably at risk right now. I mean... we should be concerned about how we will do science if PubMed is no longer a thing....

EasyPubPlot - Super simple amazing customizable plots in like 90 seconds!

 

Where was this when I was stinking up the NIH submission system with my crappy grant applications? 

You want a 110 pt font axis? You get a 100 pt font axis! 

Introducing EasyPubPlot! 

And I'm not messing around guys. This isn't easy for someone who knows off the top of their head what version of R is on their computer. I'm not talking easy for someone who knows the difference between a .txt and a .tsv and a .csv that only opens right in Notepad++, not NotePad+ or NotePad-

I want to take the vomit inducing plots from SpectroNaut and make a plot that is not vomit inducing? 

I go here. 

https://pharmaco-omicslab.shinyapps.io/EasyPubPlot/

I put in a .csv that contains my protein, my fold change and my p-value and -BOOOM volcano plot! 

Is it going to impress your bioinformatics friend carries a shoebox of hard drives with them wherever they go? No. Because he/she/they can make a volcano plot some other magical way that I can not. 

Did I make the nicest proteomics volcano plot I've ever made in my life in about 90 seconds? HELL yeah I did and then I cranked the fonts up to 110 and turned off the legends and swapped the colors around. I DID THIS WITHOUT LEARNING ANYTHING. AT ALL. 100.00% of my brain was focused on a ridiculously good bibimbap. 

For real, it's like they took everything I need to make plots and simplified it down to just those things. It's like what the Red Elephant did for manual peptide sequencing. Or simplifying Prosit down to actually what I need from Prosit the way EL FRAGMENTADOR does. 

I've been paying for and using GraphPad for 5 years and I can't make something in it that isn't hideous without 4 hours of work. It's eventually okay, but - man does it start out dumb every time you put data in it. EasyPubPlot just makes nice plots! 

It does more stuff than Volcano, too! This is just what I'm most excited about. 

The heatmap feature is nice, but if you need a heatmap the Broad's Morpheus toolkit is probably a little nicer and it has embedded statistics. But if you just want to make a nice publication ready bubble plot or box plot - the tutorials are ON POINT. 

Is there a problem? 

The pub is short for "publish" but could you make a publication ready volcano plot at a pub? There is only one way to find out, probably. 

Too lazy to read? Navigate NCI resources with DrBioRight 2.0!

 

Just like everyone else, I'm absolutely fucking thrilled to see Artificial Intelligence showing up all over the place. A lot of commercial websites now have slow new AI functions that are essentially "ctrl+f" in your web browser - the kind of improvement is assume Elon Musk and his fans are most excited about. 

(Another amazing new success from Mr. Musk this week. Add that with the fact that they're  

now leasing office space for HHS bioinformaticians to log into AWS to do their work...cause that's a cost savings over them logging into AWS from spaces that cost the HHS $0! .... Though my favorite thing this week is the stated attempt to put AI into all government databases.. ..which...run...on.. .COBOL.... if you aren't familiar, when I took Fortran 77 in highschool (the 77 was the year that version of the language was finalized - which is a good summary of Appalachian high schools in the 90s) my teacher would make fun of COBOL being old and useless. Totally going to seamlessly integrate with AI. No worries there. 

While I'm ranting you can absolutely buy this at BestBuy right now. Just in case your "smart washer" from 2018 that is the single worst appliance you've ever owned in your entire life needs downgraded for 5x what a non-smart washer/drier that works costs! 

With that lead in! 

Introducing Dr BioRight 2.0! A large language model to search NCI data! 

If you're thinking something like "oh no, did peer reviewers see the words 'large language models and multi-omics' in the abstract and just accept it without ever looking at it?" you're probably a cynical jerk. Geez.... I mean, that's exactly what I thought, and that's why I spent my morning trying to see if this LLM could actually do things (other than generate "Internal Server" errors, which, to be fair, it generates an awful lot of) 

Short summary? There is some value here, I think. Particularly if you're not a big fan of reading, but you're patient enough to ask an algorithm to dig through some results for you, and you aren't patient enough to dig through said results yourself. 

For a bad example - I went to a pancreatic cancer dataset and I asked Dr. BioRight how often KRAS mutations show up in the cohort. Since it's the most mutated gene in that dataset, I figured it would skim the abstract and give me the answer. It didn't. ChatGPT400 or whatever it's called now will do that for you, that's not why you want to use this tool. 

I got frustrated, saved this blog post in my "do not push the publish button, Ben" folder and moved on. I went back because it popped up in my newsfeed and then I grabbed a random dataset and started asking it stuff about the source data in the study.

 And this is where things get really interesting. In this study with a 65 patient cohort I just asked what the most commonly mutated genes were. It appears to generate R commands and run them?  

That's actually super legit! Slow....but it really sounds impressive. Now - I'm not checking to see if this is actually right, but you can see where this might be an asset.

You don't go to an AI to make art for you because you know how to make art. You don't ask ChatGPT to write something for you if you want it written well and accurately. And you don't want to ask Dr. BioRight 2.0 about a cancer study that you reanalyzed yourself and contributed to a paper on. But if you are interested in skimming publicly available data for stuff you want to dig into depth on, it is faster than pulling the manuscript source data, figuring out which table is the one you want, and looking for it yourself. And that can probably be of use at times. 

Multi-omics of coal miner's exposed to 10+ years of coal dust....

 

I was really torn on using the above .gif because I come from a long line of people who have died either in coal mines or thanks to what they inhaled down there. Unfortunately, all of my cousins that are under the ground somewhere in WV bathe in the political disinformation kool-aid to the point that I can't even talk to them, so we're going to have some fun with this important new study! 

If you think being a coal miner in West Virginia is dangerous just because the governor of the state made his billions of dollars running the least safe mines in the USA and never paying his legal fees - you'd be right. But it's a super dangerous thing to do anywhere in the world and there is always some piece of dog shit who is willing to save $1 by making it a little less safe for you. 

This study found a cohort with the following requirements - you had to be over 40 and you had to have spent more than 10 years breathing coal dust. How'd they know they weren't lying about being down there 10 years? Well...most of them had easily diagnose-able pneumoconiosis. Which, yes, I had to copy and paste. They did have to exclude people who had lung cancer... ugh...so if you know Zoolander, maybe the funniest movie that ever happened that doesn't have Will Ferrell as the lead, you know what .gif I should use next and I probably shouldn't.

They didn't go after plasma - they went after BALF (no, I can't spell it) which is a fluid in the lung that is not fun to get from people. I hope they compensated these volunteers very very well. 

They FASP'ed the BALF (man, if you are thinking of FASP'ing human samples, I'll legit send you S-traps they're like $5 a sample, particularly if you cut someone's lung while they were under anaesthesia) for proteomics using a TIMSTOF Pro and they did the metabolomics on an Exploris 120. Everything looks pretty standard. MaxQuant for the proteomics data and MetaboAnalyst for metabolomics. 

I can't exactly follow where the ELISA came in here in the study - like what materials and why when they had proteomics on the BALF. I figured we'd be doing ELISA on cool blood targets, and maybe that is what they did? Oh - they did ELISA on some BALFs from these patients and an unanalyzed (by LCMS cohort). I think.

Hey - I wonder if the proteomic or metabolomic alterations of breathing coal dust for 10 years is obvious? Oh.

Yeah....it might be obvious.....

I'm not sure what the authors are planning to do with these data. I mean, you could probably put together a powerpoint presentation on - inhaling coal dust for 10 years alters your lungs on a basic molecular level? 

Completely unrelated - 

MaxQuant updates for isobaric quan - TMTPro/MBR and no reference channels(?)!!

 

Oh yeah! Can't wait to try this out! 

Most interesting, maybe is that in the application of isobaric match between runs and new stats with/without reference normalization - it looks like pooled reference channels between plexes may not be necessary. The link above is my kluged together solution for matching between TMT plexes without reference/pools using high abundance "boring" proteins and I largely put that post there so I could find those same data and quickly compare the two. 

Something I hoped to get a good answer to from the paper is whether the TMT MBR was working well for non-Orbitrap data. I used it for a relatively small TIMSTOF TMT set from early 2021 and it did appear to decrease missing values, but n=1 and I could have just got one of my settings off between runs 1 and 2, but it could be a gamechanger for some of the stuff we're working on. You want to hear that it is working before you ask MaxQuant to load up 600 Bruker .d files, though, but sometimes you just have to YOLO and apologize to the other users of that PC later. 

US HUPO Video Contest Award Winners are live!

HAHAHAHAHHAHAHA! 

We didn't have the time to do an unveiling of the winners of this year's US HUPO video competition. 

But now they're live! 

The Tragedy of The Quadrupole - from Mobilion

 and Omics Family Therapy from UC Davis!  

Spend 3 days doing a pulldown and analyze it in 7 minutes of LCMS time!

 

Okay - I won't lie - 7 minutes to fully characterize an immuno-affinity enrichment experiment (IP-MS or AE-MS or whatever) is pretty cool.

I do have to wonder if that is the most effective use of resources. And - yes - I'd have the same question if this was a TIMSTOF Ultra or other super high end instruments, but only because the pull-downs I've seen seem to take students 2 days to do a pair. 

Presumably...you know....the instrument isn't just sitting there after it does it's 14 minutes of work....

It is super cool that the bottlenecks would fall back on sample prep and maybe through robotics and stuff you could find antibodies to do pull-downs on all the things. 

But you do have to wonder if maybe it is a tiny bit of overkill? Like...could you do this on an Exploris 240 for 1/4 the price and just run a 30 minute gradient instead of a 7min? It's tough to imagine a scenario where this is a superb use of resources. 

However - if you're in a busy core facility and you've got 40 people who do IPs and you the 3rd Thursday of every month is IP day (yay!) knocking out 40 pairs of IPs in triplicate in 

40 x 2 x 3 x 7 / 60 in 28 hours would be pretty great, particularly if you're paid by sample injection for 240 in about a day's time. 

The Harvard thing where they've done just an absurd number of pull-downs is one of my favorite resources ever (and completely under-utilized by the scientific community) - PlexBio or something? And the thought that you could do that label free between now and ABRF (if you had the samples) is super compelling. 

The Proteomics Show is officially top 3 proteomics podcasts on a service that lists...one!

 

This guy wrote me this morning from something I'd never heard of to tell me the good news! Out of a field of 1 proteomics podcasts - THE Proteomics Show had made THE TOP 3! 

On further investigation, the show is actually ranked #1 out of all 1 proteomics podcasts! 

Huge thank you to whatever this thing is called and whoever votes and for not picking up the much more professional proteomics podcasts by Nautilus and O-Link. 

In the latter case, we were just on that show! 

As always, thank you US HUPO for paying our editing fees and annual software subscription and for our new microphones. Well....I got an open box/refurbished microphone and saved US HUPO $15 and that is starting to seem like a mistake, as you might catch from the show from time to time. I'll be super bummed if I have to buy a new one for $90 later and use this glitchy one for parts. 

Most of all, thank you to everyone who replies to emails when we invite you on the show! Without 60 amazing and often famous guests over the last few years, it would just be Ben and I rambling at each other and we can't really get away with that more than once each season. The magic of the show is that we have smart guests saying smart things and then Ben and I dilute it down with our mouths and brains. 

Fingers crossed we just proposed something completely different for a Season 7! 

Also for not podcast people - your phone probably has a podcast app and you can just open it and type "proteomics" and ...that's probably us.... or Amazon music or Spotify or other things. 

Phosphoproteomic analysis of different CHO - Antibody producing cell lines!

 

I don't manufacture antibodies, and I honestly only have a loose understanding of the process. If I really like you, or your project I've got a 7 minute gradient, really expensive Thermo column and even more expensive high temperature heater necessary to make the column work properly to knock out standards that look like this 

...and data from your mAB that look more like this ...

Which is clearly your protein, cause my standard looks a great. It isn't that it was selected from a broad range of possible proteins to find the one that was absolutely the most amenable to mass spectrometry, or at a level of clean that is completely impossible in a production environment.

I know that CHO cells (Canadian Hippo Organoid - name suggested by a 4 year old, it actually might be something that sounds far more disturbing) cells are often used for production. 

Want to know more about the process? Check out this new paper! 

It turns out that there are multiple lines of CHO cells and some are good at making some antibodies and not others. This group investigated some cell lines that were good and less good at making particular proteins to see what the difference is. I'm not entirely clear on why they went straight to phosphoproteomics, but US HUPO 2025 was amazing but my work backlog wasn't and my brain is still toasty from digging myself out. 

Turns out, though, that they are really excited by the patterns of phosphoproteomic alterations. From a production standpoint I'm not sure if the idea is that they could leverage some cell types and selected conditions to force advantageous production conditions - or if the idea is better applied to selecting the right cell lines for product production based on these information. Either way - antibody based drugs aren't going anywhere and it's cool to see proteomics addressing riddles beyond host cell contamination! 

Proteomics reveals new cell phenotypes in a marfan model!

 

Geez. Cedar Sinai is starting to show off what they've been working on the last couple of years -and this is another complete gem. 

True story - I completely changed my plans for my first ever solo R01 submission (which I submitted whether or not anyone is there to read it) based on the paper they just dropped in MCP. 

I should really really get back to reviewing large instrument proposals cause they're like - due - soon, again - I'm already several hundred pages of meticulous writing (not mine, other people) and I'm going to turn them in when I said I would regardless of my suspicions I'll receive some email bounces. 

This preprint is stunning, though. And...intimidating....now I need to figure out how to make plots that look this fucking great? Yikes!

Wait. Clearly I'm sleepy and the foucs here is obviously that a team of top notch people put a shitload of work into a terrifying (and relatively rare) disease like Marfan? So so so so so so so cool. 

So cool that I didn't put the word "single" in the title so that it won't lower traffic to the post. For real, it is still a thing. 

FFPE vs Fresh frozen tissues - very few differences?!?!?

 

I'll probably eventually convert my pile of notes from #USHUPO2025 into ...something.... but I need to knock out some papers that you should read (and leave them on this site so I can find them later) 

Let's start with this "WAIT. THAT DOESN'T MAKE ANY SENSE, BUT IT'S AMAZING!" study

This little team answered a question that I'd assumed I wouldn't want to know the answer to - FFPE vs frozen tissues - obviously there are nasty differences that will make everything impossible to compare, right?

They did a crapload of samples for whole proteome and phosph-proteome at depth and found that - actually, the proteome seems shockingly well preserved in both cases. Fresh frozen tissues have some advantages in phosphosite IDs, but I'm shocked by how similar the results are. I have honestly always figured that FFPE would be something you'd do because it's just simply easier and way cheaper to store these samples. But if there is truly limited loss in proteome integrity at scale? Priceless. 

PSCS - A tool for sharing reproducible single cell analyses!

 

There are tools out there for building pipelines and sharing single cell -omics data analyses....but I sure ain't seen one from a proteomics lab yet! 

Check out SCMS. I would have missed it due to all the #USHUPO2025 fun, but it was a key feature of Aleksandra Binek's excellent poster and she set me up with links to the preprint and the online server! 

If you've got some SCP data sitting around you're trying to make sense of, there are written and video tutorials already up and you just need an academic email address to register and get in and mess around here. Just don't overwhelm it while I'm trying to get the hang of it. 

Finally! A comparison of different modalities for plasma proteomics!

 

THIS is the study that I've been waiting for. I don't know how it wasn't blowing up the interwebs when it dropped, but here it is. 

MS2Bac - Explore, identify and classify bacteria across hundreds of species!

 

Despite the fact that bacteria have little genomes (1,500 protein coding regions for the tiny ones and 6,000 for the big complex ones like Pseudomonas?) not even all of the model bacteria have been explored really in depth by proteomics. 

This study goes a whole lot of different directions with some great findings pretty much every way the authors point it. 

The data from bacteria were acquired on a nanoflow Orbitrap Eclipse and a microflow Exploris (50uL/min @90min). There is a helpful table that tells you which experiments were performed on what and when. The authors also do classical 16s rRNA sequencing stuff on dairy products infected with unknown bacteria and acquire a bunch of data, including proteomics of over 300 bacterial species with data dependent analysis.

MS2Bac is the software and, suprisingly, it leans on spectral counts as a proxy for quantification. For younger people, spectral counts relies on the fact that the number of peptide spectral matches for each protein is an indirect function of that protein's relative abundance. Imagine bacteria A has 5,000 copies of EtfA protein and bacteria B has 100,000 copies. When you do label free proteomics of both, the bacteria B will probably have both higher sequence coverage of EtfA and a higher number of repeated fragmentation of the EtfA peptides. We used to do this because we didn't know how to do direct quantification. The spectral counting approach is also generally handicapped on high resolution instruments like the ones used here because an Orbitrap is smart enough to recognize that the +2 and +3 verions of peptide X are the same thing and will only fragment the most abundant ones, reducing the proxy for linear dynamic range of spectral counting techniques. Given the fact that there is a 150 Hz ion trap on the market right now, I suspect that we'll see more spectral counting.

I don't think many microbiologists will be surprised to know that you can tell bacteria apart pretty darned well by LCMS. You can put colonies of bacteria on a plate and hit them with some MALDI pulses and the spectra coming off is a decent enough signature that you can choose the next clinical intervention. There is a big difference between that and a tryptic digestion, however, and this is a bunch of bacteria that look different even when they are closely related. If you are a microbiologist, I don't think you'd be sad to have access to 300+ bacterial proteomes, either! 

Top down proteomics by EAD!

While I am skeptical of the currently accepted machanism for how those magnets induce electron activated dissociation (EAD), it's a super cool new technology. ETD but with basically no reaction time and no scary reagents?  Wins all around.

In this newest application of the technology these authors apply it to top down proteomics

They mix up some commercial histone standards and show they can resolve and identify their PTM sites, then they do some bigger proteins (heavy chain might be the biggest protein they try in the paper) and then also resolve some intact glycosylated proteoforms.

I thought this was a neat trick for overriding the default limiter for the ZenoPulsing (for the shotgun proteomics methods it was designed for you're more worried about saturating the detectors than what happens if your signal is a little lower, which is a very different problem than for intact proteins that really really don't want to fly). 

My very first question was "cool, but what can process these files?" and they used OpenMS and Flash Deconvolv (something like that? I might have the name partially mixed up with another similar thing and I'm not going back to the paper to look it up - definitely had Flash in the title). I'm pretty sure that if you can master the OpenMS framework that thing can process basically anything and this supports that assumption.

Holy crap - US HUPO starts this week (Saturday! 2/22/25)!

 

Ummmmm.....that's this week! 

Fortunately, I only have responsibilities on Sunday (Early Career Researcher "Difficult Conversations in Science" - woo!) Monday - co-chairing the most popular and absolute best session - Tuesday....ummm ... should probably make a slide or two....

And the astrologer they use to provide weather forecasts to my telephone says that we'll only have 15 inches of snow on Thursday. Which, given current accuracy means absolutely no precipitation whatsoever and I probably won't even need a jacket! 

US HUPO sold out of the conference block of rooms, but you can still totally go. I registered last week! 

https://www.ushupoconference.org/registration

Oh yeah, and these are the 2025 US HUPO award winners! If you are a podcast person, check out the new The Proteomics Show drops! We get to talk to all of them and Professors Cristae and Fenyo are up wherever you get podcasts! 

Single cell lipidomics by nanoLC TIMSTOF!

 

Wow, am I super pumped to see this one out! 

You know all those minimum handling techniques for single cells generally means that EVERYTHING is injected into the mass spectrometer, right? 

This is a pretty minor alteration from my favorite paper of 2024 - CellenOne pulls out single cells, but instead of digesting them you drop in lipid standards and methanol and then run the TIMSTOF optimized for larger small molecules (lipids). 

They have to do optimization in both positive and negative mode, the latter I've never even tried on a TIMSTOF unless I was doing MALDI, and then the entire screen changes. Fortunately, we don't have to do a darned thing, we'll just steal this method entirely. 

Immunopeptidomics on low cell numbers without crappy antibodies!

 

Big shoutout to Anthony for sending this preprint over! 

We spend a lot of time talking about how much specific antibodies can hurt science. We've been talking about it for as long as I've been able to have somewhat intelligent conversations about science. 

What hasn't gotten as much attention is how terrible "pan" antibodies are. I see papers all the time where they still do phospho-tyrosine enrichment with the 4G10 antibody- and I first used that thing 15+ years ago it wasn't very good then, and I ask at conferences all the time if it's improved and I don't think it has. 

You know what is way way worse than 4G10? The absolute pile of poo that is used to pull down the conserved regions of HLA (MHC) proteins. They are clearly biased toward specific epitopes, if you want them to work well you have to slowly rock the antibodies with 4 kilograms of cell lysate, and it helps if you do it a few times and the same procedure carried out by 2 people following identical protocols get very different results. (This last part I saw at a meeting in 2023 but if the presenting author has published those results -she was in pharma R&D - I don't see it on a quick search). 

CAN WE DO IMMUNOPEPTIDOMICS WITHOUT GARBAGE ANTIBODIES? 

No joke, this is worth taking a look at. 

Mild acid elution has been around forever. Heck, the best way to get a zillion "immunopeptides" identified is to use a not-so-mild elution that will break open the cells and release the endogenous internal peptides and then to turn off your peptide FDR --> boom! a zillion peptides. To be clear, I'm only recommending this strategy because there are at least a couple "CRO"s out there that will do something similar and charge you $10k/sample for completely and totally useless garbage results that hurt our field while making people I can't tell are unscrupulous, dumb, or both a whole lot of money. 

Actual mild acid elution, though, makes a lot of sense if you can fine tune it in. Now, there is some controversy out there about whether that has a bias. It makes sense that from the shape of these little presentation "binding" pockets that 3D structural stuff would weigh heavily in  on what pH conditions would cause an effective mishaping and release of the presented antigen. (Stole this image from here) 

It's actually helpful to this rant because MHC II (the ones that are way easier to analyze by LCMS because they're bigger) is generally thought to be much more resistant to acid antigen release stuff. Although you'll find people who will contest this one as well.

What was I typing about when I should have been working on a proposal on my desk....oh yeah! This paper! 

So these authors do a very standard crappy antibody bulk pulldown from your typical 800 billion cells or whatever and they look at it on a TIMSTOF Pro. Then they do some comparisons to their acid elution thing which is helpfully contained within a simple little tip that then goes right on the EvoTip (wooo!) to get down to the 50,000 cell and below they do go to a TIMSTOF SCP and run the nanoflow (Whisper) methods on the EvoSep. They stick to DDA with fine tuned polygon IMS/m/z windows for immunopeptides (rather than just no polygon, which some people use for endogenous peptides on the TIMSTOFs). They also compare this to the super small window DIA stuff that they recently developed. That was called something confusingly similar to diaPASEF methods by other groups, but I think they decided to call it Thunder, because if there is anything the proteomics field needs more than confusing new nomenclature is is absolutely and without a doubt the same method from the same group with different names. 

Seriously cool method and I'm absolutely going to be trying it myself this spring. Probably step by step, with the exception that I've got an Ultra2 arriving in 25 days! 

Was your LC column provider acquired? Try a new vendor out, save some $$ and get a free column!

A while back my nanoLC column provider was acquired. There are always growing pains with those things, particularly when production is shifted to a new location or continent. 

I switched to a nice small company with super friendly people - 

and some techbro bought that one too! I've been part of a couple of failed startups. I know a techbro when I run into one. I was so enamored that I immediately got a column bomb and dug out 15 year old notes on how to pack columns. As you might be aware, I have a blind spot in my right eye from accidentally ramming a rusty screwdriver into it during a bathroom remodel in 2020 that still isn't done. Maybe packing columns isn't something I should be doing, but desperate times...

Wait - Premium columns at affordable prices? For real. Check this out.  

I'm impatiently waiting on my first ever Ben owned mass spectrometer. A TIMSTOF Ultra2 with EvoSep. I ain't gonna blow it up with some dude's first column he ever packed. I contacted the owner and - dude's been making nanoLC columns for like 20+ years. He's US based (so...tarriffs shouldn't blow up my budget) and if you order a column they'll send you a second one for free. The columns are also lower priced than anyone else's to start with. 

The two columns I use the most are on that bottom row. These are close matches for the EvoSep recommended part EV1137 and EV1109 which I used to get the equivalent of from the newly acquired company. Stay tuned for a head to head when my stuff arrives in a few weeks! 

Again - Disclaimers over there --> PremierLCMS didn't give me anything for this, except they're sending me a free column (and Kerry is going to help me find an ideal matrix for a mixed mode application we're developing) and they'll send you one too if you move fast 

Multi-plasma plasma proteomics study will make you want to invest in O-link!

 

Look at these nerdy Europeans and their multi-institute studies! This one is on the achilles heel of mass spectrometry based proteomics - non-depleted plasma! 

It looks to me like they prepped plasma and sent it out and everyone could basically run it however they wanted on whatever they wanted - and 

....400 proteins is the high water mark.... now, this isn't enough for me to stop running plasma on mass specs, but it is enough for me to go across the hall later and ask how that O-link thingamabob we have here works! 

Formaldehyde fixed single cells maintain proteomic alterations!???!

 

Leaving this link here so I can find it. There is a point sometime when an ASAP JPR paper appears google searchable and may not make the ASAP page tab on the JPR website. Thanks @cportgrrl.bsky.social for helping me find it! 

This straight up surprised the fuck out of me, btw. I assume when you dump in formaldehyde for any length of time you just stuck all your lysines together and globbed up the nucleotides and left oxidative secondary changes all over the place.

This does not appear to be the case! The authors take some HeLa cells and hit them with a DNA chelator (5-FU does that, I think, don't quote me, haven't used in years) and some other drug and looked at fixed vs. unfixed cells. You'll see some huge numbers in the early figures. 8,000 protein groups. That is from 1,000 cells digested in the top of an EvoTip at a time. 

Protein groups drop a little with the brief formaldehyde treatment, but the differential proteins upon drug treatment look about the same - in some cases better - with formaldehyde rather than no treatment.

Does that mean you should quick formaldehyde your single cells? Maybe not - but maybe that would stabilize them if you can't analyze them immediately? Maybe that helps with shipping? 

MORE IMPORTANTLY! Of the tens of thousands of samples I can access through my super cool new blanket IRB and hospital repositories at my university - about 80% of them are formalin fixed! I'm getting very bad ideas thanks to this really cool paper. I need to reproduce it first before I go spend money and time on said bad ideas, but - hot dog, yo, this could be huge.