Your local air (barometric) pressure changes your TIMS pressure! How to fix!

 

Man, this is so freaking smart and inciteful - and comes with suggestions for how to fix it! 

Now, the rumor is that the reviewers put them through the ringer on this to make them prove this wasn't something isolated to wherever the heck Heidelberg is. So they extracted data from files in ProteomeXchange from other locations and since the Bruker crazy .sqlite database format keeps information on freaking everything - they could pull weather reports and show the same stuff happens in Boston. 

Totally worth keeping in mind, because if you're doing something like the last paper on the blog where you do little tiny 1/k0 isolation windows - but...those...change....cause your air pressure drops.... that little window isn't too great all of a sudden. If you're running the "long gradient default methods" where you're scanning from 0.6 - 1.6 it probably doesn't make a difference, you're already fragmenting tons of dumb shit anyway (like PTMs and short peptides and long ones) that you won't identify. Super duper smart paper. 

Single cell proteomics identifies new (sub) cell-types in the human heart!

 

WOOOOOOOOOO! Man, this paper is a lot of fun to read. I can't spend a ton of time typing here. I've got tons and tons to type elsewhere, but it's just in time to be cited in what I'm writing. 

Why would you read it? 

1) This demonstrates that terminal cell types that we like to categorize as "cardiomyocytes" that don't really divide and more or less hang around the full extent of your life...

- have different proteomes! Does that make them a different cell type? Does that make them a "sub-cell-type"? I'm not the person to answer that question, but it sure does throw questions in a lot of places.

Really interesting methods in play here as well. Dual Trap Single Column loading with short little columns allowing some really fast cell-to-cell throughput? Pretty statistics stolen directly from the single cell RNASeq people? Not the highest coverage you've ever seen per cell, but a demonstration that if you do a bunch of them you can find really cool stuff to look at. (Maybe 800 proteins/cell average?) 

Importantly, they didn't just get cells from hearts they also took some stem cell things and made those into cardiomyocytes and studied how those were the same (and not the same) as the cells from the actual hearts. 

Minor gripe - the study is definitely a bunch of different cells from different places (5 types of IPSC derived cells plus some cells from ACTUAL HUMAN BEING ACTUAL HEARTS) but the abstract and conclusion bill it as a very large study. It is 1326 cells. Which is a lot. That's as big as the SCOPE2 study from 2020(?) and we all sort of aim to do more cells than that paper. There are probably 8 studies that have done that many cells or more? Was it a crapload of work? Hell yeah it was, but it'll disappoint your scSeq friends. 😇

Change your TIMS settings across your gradient for a crapload more IDs!

 

I legitimately thought this was going to be one of those silly mass spec things we all do where we tinker with settings and get 5% more IDs at the cost of reproducibility.

It is probably the second thing, but the improvement level is worth taking note of. 

The paper is open access so check it out, but a TIMSTOF Pro going from 5,000 some protein groups to 7,000 protein groups (? hard to see on the bar plot? by using 10 different 1/k0 isolation settings across the gradient vs the default instrument settings? That's a surprising improvement. Worth noting that the gains aren't as impressive at shorter chromatography - you see the biggest gains at the longest relative gradients. 

Downside here - like just about every TIMSTOF paper out there - I don't think that I could walk up to the TIMSTOF Pro2 in our core and reproduce this method exactly from the method as described in the paper. Thats a bigger problem for addressing later. 

Need some enthusiasm today? Check out MODE!

 

Okay...so not to distract you guys from doom scrolling the news or the "news" if you're in the US where the filters on what we're seeing are...blatantly obstructive at a level we've maybe never experienced. 

Need some youthful enthusiasm? Start here. 

But actually go here. https://map.emsl.pnnl.gov/app/pmart-classic - you will need to register with PNNL. 

But then check out the short tutorial videos on how to use it. I had them playing in the background while I puzzled through making a plot or two. The narrator of the video seems to be really having a great time walking you through these tools. It carried me over enough to write a blog post about a really nice little toolkit on a dark and cold day. Maybe it'll inspire you as well! 

Trump stopped all NIH study sections??

 

I've been moving and writing and just got internet in our new apartment a couple hours ago and - I'm leaving this link here. 

I'm going to take some time to permanently block on my browser any "news" site that doesn't have this on the front page. Why do you think that these right wing fanatics think that the US has had a technological lead on most countries for decades? I seriously wonder, because I think it's the fact our government has invested in unbiased research and that every dollar has translated into hundreds of dollars eventually. 

Scary times. Stay safe and positive out there!! 

Single cell ICP-MS without fixation using a microdroplet source!

 

Man, sometimes y'all send me papers and totally ruin what was otherwise a really focused run or something. This paper is certainly the case, because it is a fix for something I had no idea was an issue. 

We've had ICP-MS of single cells for at least a decade. Maybe much longer? I don't know. There is even a commercial FACs coupled ICP-MS thing that I'd love to have. So when this paper popped up in my inbox with the suggestion that it is a big deal I thought I was just being sabotaged. 

Did you know that when we do ICP-MS of mammalian cells they're fixed in formaldehyde or something first? It's true, I checked method sections of a couple papers like this one. 

This may not always be a bad thing, particularly for cell surface markers you're tagging with an antibody with a metal chelated to it. However, if I wanted to measure how much of a metal containing drug was in my cells, I think I'd be nervous about the effects dumpinng formaldehyde on them would cause versus the original conditions. Even better, while this is sort of in-house developed, it was done using commercially available tools and the instructions are clear enough that it makes me feel confident that you could put it together. 

They used a really nice tool I've never actually seen in person - an ICP (Inductively Coupled Plasma, in this case, not the HPLC thing) triple quadrupole (Agilent 8900, I think). I don't see why the second and third quads were necessary. This matters because the last ICP single quad I bought more or less in the HPLC price range. I looked for a quote but couldn't come up with it. I did find an ICP-QQQ quote and that was around what you'd expect for a good LC-QQQ, somewhere in the $300k range and I couldn't come up with a great reason for why I'd want the other 2 quads so no one would pay more than 2x for them for me. 

That's a lot of words for - this appears to be something that could be assembled relatively easily at a not entirely insane pricepoint. 

Reposting - advanced mass spec operations -Exploris and TT Ultra - 3 hour video!

 I was looking for this and realized I'd put it with a bunch of other new videos I'd found.

This is from last summer's Northeastern University /Parallel Triangle Institute hands on single cell proteomics day.

You get some super cool insight into running an Exploris with MaxQuant.Live AND expert tips on the TIMSTOF Ultra operations. Great great resource. https://www.youtube.com/watch?v=mMrujxr3JFw

Second crazy high depth single cell paper dropped same day from another lab!

 

The preprint of this paper is on the blog here somewhere I think, but this paper has progressed quite a bit since the initial posting. 

I'm pretty sure the preprint used just standard "Whisper" 40 and 80SPD. It's been a while, but that's my recollection. Those methods are somewhere in the 100nL/min elution range, though if you watch the EvoSep monitor closely it sort of goes to 50nL/min but maybe just for equilibration purposes. This appears to feature the new EvoSep WhisperZoom methods which are 200nL/min, or what you might consider ...the default method on your instrument for nanoflow liquid chromatography... 

The authors also decided to add a completely unnecessary new name to their method because the one thing we need while mass spectrometry based proteomics is rapidly being displaced by "next gen" technologies is more confusing nomenclature. What's Chip-Tip? It's using a CellenOne to load a 96 well plate that can be loaded into EvoTips. It's a vendor provided protocol that you can purchase right now, and name whatever you want. You put down like $1.7M for this whole workflow if your vendors really really like you, you can call it whatever silly thing you want to. Snarky annoyance aside this is both interesting and useful information --

In this iteration, which I do not have the time to read fully (disclaimer - I'm skimming and running an excitingly high fever) - they do employ the WhisperZoom methods which now go up to 120SPD. That's one single cell every 12 minutes! On a standard commercial system! I've got some of those small Whisper columns here and I'm taking one apart when the emitter inevitably clogs to figure out how much C-18 is actually in it, but it's probably like 5cm. 

How's the chromatography on a 5cm column on an Orbitrap based system running 4amu Asstral windows? Here are some of my favorite examples from the nifty new DIA-NN viewer! 

Sweet! You don't get much more gaussian than a literal f'ing triangle! 

Here's a Rhombus! Is that a rhombus? Maybe it's the fever but I initally thought Wesley Snipes would carry something with an end this shape vampire hunting. 

There are good ones here, though. I was just jumping through one file after thinking - wow - that seems like a slow method for a really really fast chromatography gradient. Here's a really good one. 

For real - an astouding protein coverage for a single actual cell! 

I would like to point out that there are other Asstral single cell preprints out there, Schoof lab has some amazing stuff on biorxiv coming soon, I'm sure. They utilize  more traditional DIA isolation schemes to get amazing chromatography out of the instrument. The Asstral's ability to accumulate DIA data with tiny windows is super super cool - but even at those speeds, it isn't a short cycle time and might not match well with super short chromatography! 

Up to 5,300 proteins per actual single cell!?!?

 

Holy shit. I'm way too busy today to spend the time on it that it deserves, but I'll set some files to download until I do...

Up front - I'm biased. IMP is one of my favorite places in the world and this study features both friends and a personal hero or two, but this is a sick paper.

I do very much appreciate how clear the authors are here about cell size, etc., they don't average 5,300 proteins per A549 cell, but they do get that on the largest ones with the highest relative protein content. 

There is a lot to unpack here. They start with peptide dilutions, but then move to multi- proteome mixture to make sure that what they're seeing at low loads are relevant quantitatively. All throughout they are transparent about carryover and other factors that mean they are seeing proteins in their "blanks". 

Y'all, it wasn't that long ago when I wasn't getting 5,300 protein groups single shot proteomics on A549 cell lysates regardless of how many cells I started with. The fact a big A549 cell could give you anywhere near those numbers is just absurd. 

Thermo scores 600,000 sample proteomics study! With O-Linking!

 

600,000 freaking samples? FOR PROTEOMICS! 

The goal is a depth of 5,400 proteins/sample. 

Quick reminder that when you're using this platform you are PROBING for 5,400 samples. That doesn't mean you detect them all. Obviously they might just not be there. Or they might be below your linear dynamic range. But you tried to measure 5,400. 

I'm unclear from the press release what these samples are. All plasma? In which case - 5,400 is going to change the game entirely. 

The human blood atlas (thanks Ben, for the link to this updated page) only lists 4,200 proteins detected by mass spectrometry!