More crazy SCIEX 8600 proteomics data, with good chromatography AND public data!

 

Okay, so this is a vendor application note, but it's not written by company scientists. In fact, it appears to be the next edition of this sub-radar preprint from earlier in the year. 

I've had this preprint open on my desktop for a while in tab 713, because it's the first time we've seen Astral vs Ultra vs 8600 (and 7600+).

As it said on our US HUPO poster this year "vendor instrument comparisons are boring" or "...because Ben (Orsburn) said vendor instrument comparisons are boring..." maybe that was it. In the preprint the authors completely and totally avoid any point where they set up 50ng of the human/yeast/ecoli mixture and the same relative chromatography and compare the results from these same instruments. That's not the point. The point is a current generation publicly available dataset to compare your proteomics algorithms on all the new shit out there.

The 8600 looks really nice whenever you see peptide IDs or quan, but if you also saw this back in February and were a big enough nerd to make it to page 17 of the Supplemental information you'll see this - 

The Astral and Ultra2 chromatography looks pretty solid. Particularly, the refined Astral gradient vs the Ultra. ...and...the chromatography on the SCIEX's....looks like its having some issues..... And that was the problem when I had the 7600. I had the Waters HPLC and it really just didn't work properly all together. 

Why this app note is interesting is that they seem to have refined those chromatography issues..... And this currently budget instrument seems like the next generation instrument that it might truly be. Look, y'all run the HYE digest all the time. I see it in the literature all over the place. I personally prefer just a human + E.coli spike. I think adding 5k new proteins to human is a little bit excessive, but I've looked at a lot of HYE digest data. 

What I haven't seen is almost 11,000 protein groups in 15 f'ing minutes. At 50 nanograms? Maybe the Astral preprint from the Thermo marketing unit in Denmark has come close? I think they loaded more and ran longer. Given what the 11 minute EvoSep data from the preprint looks like in DIA-NN, despite what appears to be something wrong with the HPLC, it doesn't seem all that unlikely. People out there with the >$1M instruments should be very nervous about this thing. I did get an academic quote for one for $510k - with $25k/year service!! 

The 2026 list of techniques for deep plasma proteomics by LCMS?

 

This post is probably not actually about this great new review. I'm just leaving a link about it here because I'm reasonably sure one of the peer reviewers put this team through the wringer. Is that a thing? Why would it be related? I don't know, but it's a great and valuable review! 👮

This post is about ALLLLLLLL the ways we can now go way way beyond those top 300-800 most abundant proteins in human circulating blood liquids. Definitely not all, I can't keep up! Edit: Definitely not all. Geez. The last paper cited below mentions 4 other ones! Okay, maybe later edits will mean adding to this and not just correcting. 

For anyone who has never ran human plasma or serum and is excited to do this by LCMS, do I have some crazy stuff to tell you. If you put that on an S-Trap or something else where you're used to getting 4,000 - 10,000 proteins, you'll probably get about 500 protein groups on your first try. If you've got today's most badass $1M thingamabob you might get 1,000. Maybe. Probably 800 is your max. And that's it. Shameless plug, for the Equalizer plasma proteomic standard, we had a poster at US HUPO where 8 labs contributed data. An Astral Zoom running a 45 minute gradient won't get you 2x more proteins (620) than an LTQ Velos from 2011. The dynamic range is just 

for mass spectrometry. Thanks to the Great Meme Reset I can use this one for the first time in 8 years!

However...if you use new magic techniques that not one single person on planet earth can explain to any other person without the made up concept of a "protein corona" (I blame the Broad for this made up terminology -it's not worse than "the mobile proton model"), you can now dig deeper following chemical or physical enrichments.

Let's count the ways! I'm currently aware of these - 

SEER Proteograph

SP3

Ceres

Mag-Net

PCA and PCA-N 

...which is already a lot....

Geez, and there is a new one I don't think anyone has published on yet. P2 from Biognosys. I may lump it in with SP3 based below until I find out I'm totally wrong. 

Let's inaccurately compare all of them! 

SEER - possibly the originator of this whole field, however, it appears that most attempts to claim this legally as something they've developed haven't gone super well. Some may still be in progress, but at least 2 have been initially thrown out of court.

Pros - probably the best data out of anything we've seen. It is one thing if 50 of the best mass spectrometrists in the country at Cedars Sinai can get 6,000 proteins in plasma. However, I'm hearing that anyone who buys the Proteograph system and reagents for their own lab and runs it is getting comparable results.

Cons - Expensive. O-Link expensive? No. Illumina Protein Prep expensive? Absolutely not. Is the data far more valuable than either of those technologies? Yes, and particulary in the case of the Illumina solution. In my opinion, there is no greater misuse of funds in 2026 than running the Illumina Protein Prep for more money if you could run the SEER Proteograph and get actual quantification data. The fact that O-Link is quantitative and can theoretically scale to far higher levels and eventually be cheaper is a little more defensible. However, in the context of what people are used to paying for LCMS it is a lot of money for a sample prep, and that has allowed other technologies to propagate.

SP3 based - these are a magnetic bead based prep that appears to have a solid number of users. It looks less expensive than the proteograph, but I think all 5 of the published comparisons have found it to have inferior results. It should also be noted that this prep requires that your soluble blood proteins (plasma/serum, whatever) is prepared by the clinicians or phlebotomists in a very specific way. If you DO NOT have control over this, you should very very carefully read those instructions. You can't just put any blood plasma/serum thing in it from any old biobank into this workflow and expect success. It should also be noted that I have heard from 4 separate people that sometimes you end up with magnetic beads clogging weird things in your nanoLCs. There is a moderately expensive and otherwise indestructible porcelain needle on the EvoSeps and you should have spares on hand when using these magnetic beads. 

The one I've heard most about is the PreOmics one which is called iST-BC3 (which, to be fair is the one that has the very specific limitations on what fluids it is fully compatible with) but I suspect that the new Biognosys P2 one is closely related to this. I think they're both largely or entirely owned by the same company now? It would be smart to consolidate methods if that is the ase. 

Pros - mid-tier cost? Commercial support kits still mean commercial support teams (people to complain to and get help from!) My impression is that you're talking 1/2 to 1/3 the price of a Proteograph prep. So if SEER is $300-$400 or something, this is closer to $100? I'm sure it all depends on scale, etc,.

Cons? - Might not be compatible with every type of human derived blood proteins. 

Ceres - is some sort of a porous thing amabob and they're making a splash on the scene right now by being willing to send you a free kit to get going. If you're struggling I hear you can get an applications scientist on site to walk you through it. Our first attempt with a kit involved going away from the established workflow and the results were still an improvement, but below the expectations of Ceres and I think someone is coming to help us out later this summer. Ceres got a big endorsement from one of the best proteomics core labs in the country. The total plasma proteomics workflow - including enrichment and analysis - is advertised as considerably less than an enrichment by Proteograph. That seems to indicate a very cost-effective enrichment. They aren't advertising Proteograph level depth, but expecting 3k proteins in plasma was science fiction a couple of years ago. 

Pros - Solid endorsement and good numbers, and US based support team. Sounds very affordable. Tens of dollars? 

Cons - Might not be the best numbers out there yet, and there isn't nearly as much published data for it out there. 

Mag-Net - wow, this one is confusing and controversial for one reason or another that I don't want to go into. I've heard that it might be due to how one group prepared the plasma without following the published instructions. Here is a good comparison. 

Ugh. This paper points out 3 other enrichment kits that I didn't even know about! Okay, maybe I need to add to this one later. Okay, well, that's where there is a link here. These authors (who I think are completely unrelated to the Mag-Net inventors) report 4500 protein groups from plasma at "a remarkably low total cost of just a few dollars per sample" 

Their words, not mine. 

Pros - Very low cost, depending on how you do it. One core facility said you're talking about $5 -$10 in beads depending on where you get them and scale. 

Cons - Might still be some controversy on it, but an increasing body of work suggests that it works just fine. If you do want to get some support on the workflow, this might be the central workflow endorsed by EvoSep. Maybe they'll have kits? 

Perchloric acid (PC / PCA / PCA-N) 

If you want to avoid particles entirely and just want a quick step with a weird acid or two, this is interesting. I've personally tried PCA-N and I found the results kind of -meh, but I have rarely done a great job of a new sample prep in my first attempt. PCA-N is the easier one of the two, you don't need a filter thing. You neutralize out the acid with NaOH or something and then you go right into your workflow. Perchloric acid is cheap and NaOH is virtually free at the amounts used here. It is worth noting, I think, that a group has recently reported no significant increase in the proteome when plasma was depleted of platelets with high speed centrifugation, so the method might be more controversial than you'd guess. 

Pros - Almost no cost or time addition to prep. Add acid, pellet, neutralize. Filter if you're doing the original prep (with reusable filter things I don't fully understand). 

Cons - Probably the least improvement in coverage. I think in the inventor's own studies neat plasma came in at 500-800 proteins per sample and PC and PCA-N got you to 1,000 - 1,500 protein groups. An improvement? Yes. Fast? Yes. Competitive with all the other stuff in terms of coverage? Doesn't sound like it. But a PCA-N study is coming that is 50,000 samples...so even a $20 prep kit is a lot of money. Is that $1M?? Extra on top of everything else? Ouch. So if you get some acid and NaOH from a pool store and it's virtually $0 to do that step? 1,500 proteins for no additional cost or $1M for 3,000 proteins will make a difference in a lot of studies. 

Okay, this was helpful for me to type out. And then arranging them the way I did by descending cost, which does, with possibly the exception of Mag-Net, depending on which paper you believe, seems to also lead to decreasing relative coverage. 

If there is a take-away here, every single one of these will give you more coverage than a neat digest of human plasma. You're looking at doubling proteins at an absolute minimum (probably) to getting 10x or 20x more protein groups. Every single one of them will provide quantitative data. Which is something you can not get from Illumina Protein Prep or SomaScan. If you actually want to know how much of each protein is present, then that's a no brainer, and every one is less expensive than that technology. Depending on which O-Link kit, etc., that you use, your scaling, etc., it's probably close to the cost of doing SEER, and I think that Proteograph will be - in almost every possible situation - the less expensive option than O-Link deep unbiased stuff. Sure, maybe if you're doing a million samples Thermo will cut you a break, but you still aren't getting those million samples back for a year or four, I don't care where you send it. Although new preps are rumored that might be able to get up into the 1e5 samples per year if nothing breaks and you run a prep every single calendar day. Nothing is going to get you to 1e6 without buildings full of robots. Who has that much (good) plasma anyway? 

Finding long "non-coding" RNA microproteins for diagnostics!

 

I wish I had more time to spend on this one, but I've got to get out the door.

The best I can tell, no new data was generated for the study. What it did was leverage riboSeq data in public repositories in tandem with well-curated proteomics data. Looks like CPTAC and another study of hepatocarcinoma. What they're looking for is sections of RNA that were protected by the ribosome, and are therefore likely in the process of being translated to protein.

Then they take those sequences, make a FASTA database and go back through the proteomics data looking for them. Ultimately, this approach probably has applications well beyond "just" looking for microproteins, you could leverage this to look for challenging sequence based isoform/proteoform descrimination as well. Looking for microproteins, though, opens up a whole new aspect of the proteome. Given the complete lack of useful diagnostics and drugs (only 30% of these patients respond to best in care therapies today....) microproteins is a fantastic place to look for things. I expect I'll end up reading this one more than once in the very near future. 

Someone bought Ignite for $150M dollars!

 

Yesterday I really wanted to play basketball and I walked onto a court even though there were high school kids on it. Every ice pack in our home has been in use since, so I was feeling pretty dumb. 

Wow, do I feel better now. Someone bought IGNITE Proteomics for $150,000,000.00 dollars. 

If you aren't familiar, Ignite is a reverse protein array. Their only current product that I'm aware of is a 32 protein panel that costs $2,200 per sample. I think that I might limp into work and see if my team can measure those same 32 proteins in human blood or serum with a mass spectrometer far more accurately than any protein array could. Then we can all go back to what we were doing after lunch. 

The proteomics industry is alive and well today! And investors are still out there making the same decisions! 

Read more about this topic at GenomeWeb here!

Two new preprints on differentiating single cells!

Yo, these are both SO cool. I'm very very jealous. I've been trying to find collaborators who can differentiate out some cells for me for over a year now. 

I'm lumping them together now becaus I'm still trying to dig my inbox out. 

One is cardiomyocytes! 

And they use the AIP! Apparently these can only be obtained in certain countries, btw. In Austria, they'll sell you the upgrade, but you can't get it installed. 

And this one is differentiation out to motor neurons! So cool.

If you were a total jerk you could point out that this is differentiating big cells out to massive cells so there is a lot to work with (respectively) in both cases. If you were someone who does work in liver cells - which are also gigantic and have exceptionally friendly copy number distributions (600-ish proteins make up like 85% of the hepatocyte proteome and they're probably the proteins you care about in the liver), it would be better to delete this whole paragraph. Or....try to insert a GIF from a Mac and fail again! 

Ultra-high speed high resolution LCMS shows we're still underestimating proteomic complexity!

Wow, there are so many great ASMS week preprints, I can't even read all the ones from Jenny van Eyk's lab. This one, however, needs to jump the queue, just a little.

The topic of the paper is well detailed in the title. It's a comparison of Asstral DDA vs DIA. There are some gems in it, including the statement "DDA isn't dead yet". But what is really really super interesting is the relatively low degree in overlap when they run the same samples with DIA, small window DIA and DDA methods. Which leads the authors to type out this really really cool paragraph. 

Okay...so why do I love this so much? Let's go back in time 12 years or something to this paper I ranted about a second or third time here. 

In 2014 or so, these authors used an Orbitrap Elite and ran some long (by today's standard) gradients and just counted the peaks that they saw. Using a D20 let's call it "low field" to differentiate it from the next generation of D20 Orbitraps, a system with an approximately 2 order intrascan linear dynamic range - they could count over 100,000 peptides. 

This is funny today because people are routinely identifying more than 100,000 peptides in their single shot samples, but at the time it was a benchmark. We were only fragmenting 15,000 things with DDA approaches then. It demonstrated how much further we needed to go for complete proteomes. 

Now with today's fastest stuff, these authors - on one instrument just keep finding new stuff. Just by running the same sample with a different method. It should be humbling and exciting to see just how far we still need to go to actually "-ome (ALL the things) the proteome. Or frustrating, I guess, depending on who you are. Lots and lots of gold in this preprint. 

Scone - printing tiny ESI emitters for dramatic signal increases!

Could the next advance in LCMS sensitivity actually be --- 

 -- is this the most perfect gif ever for...

This new study? Is it the most practical thing ever? I'm going to guess probably not, but the sensitivity gain math is large enough to think about it for at least a minute or two, if you haven't already. 

GPT-Rosalind! Purpose built LLM for Life Sciences!

 

If you've been on this "science" blog: 1) I apologize and 2) you might know I'm very skeptical of teaching bitcoin miners how to put words together into LLMs (I'm pretty sure this stands for lazy lama math).

However, when I saw the above plot on this site for an incredibly absurdly overvalued company and their made up metrics for the performance of their tool, I had to think about it twice. 

Please note that I did edit the plot above because you can't have an axis without units labeled. As a reviewer and editor, I feel that I applied the correct units to this plot and consider it a service to the nameless person who made it. I'd send it to them if there was a name on it. It might be billions of gallons of water and trillions of watts of electricity turned into dollar bills for the crypto bros who do this lama stuff now? But I think I probably labeled it accurately. 

Waters ASMS releases - wait. is Waters thinking about proteomics again?

 

This post is a little delayed because I needed to do a check of what I know vs what I'm allowed to talk about. 😇  I'll select a random date to reinsert this. 

What is Waters doing on a proteomics blog? I guess I mentioned them last 2 years ago when they rolled out the reflectron thing. That thing was doing 100 Hz at 100,000 resolution. Which would be insane at any other point in history but is sort of normal today. 

But they had talks on DIA proteomics this year! 

And they had a new box that, according to this weird site...

Wait. So 2x faster ...is that 200 Hz?? Or are we doing less passes to get there (which would decrease the resolution)? It's funny that this sort of ran under the radar. 

SCIEX ASMS 2026 focus - efficiency and affordability?

 

Another company at ASMS is also thinking in a new direction. SCIEX rolled out just one piece of hardware, the novus55 QQQ/QTrap system.

This cute little box is focused on being space, energy and temperature efficient, while still being screaming fast. There is an example 20 min method where something crazy like 500+ pesticides are measured. It might be 800. I forget, maybe that was transitions. Still, I've never built anything above 130 MRM targets (<200 transitions, for sure)  and that didn't fit in 20 minutes.

But here was the big takeaway. SCIEX thinks that the 8600 and 7600+ systems are ready for prime time proteomics right now. And they're ready to put their money where their...mouths...are...? Is that a saying? 

I don't think these numbers were shared with me because they're secret. Normally people make me sign things if they have secrets. I was told academic pricing on an 8600 is $500k with $25k/year academic service deals. To put that in perspective -- I have a quote on my desktop for a new TIMSTOF Flex system. It has a MALDI2, but otherwise it's the exact same system we bought in 2020. Great system and I'm still writing up papers from data from it. 

Would I rather have another TIMSTOF Flex? Or 4 (FOUR. QUATTRO.)? 8600 systems? Or 2 and a half Asstral 1 systems? I think my best Asstral 1 system quote was $1.2M without an LC and stuff. This was during my lab setup and I know that prices change all the time. Actually...

As a reminder - just a couple years ago these systems were >$1M. 

I signed up for a demo for whenever they can squeeze in some time for me.  Because scientific funding isn't getting better here.  And I literally don't even know how you fund a system over $1.2M USD? Like...how.... for real...actually asking.... I told everyone at SCIEX I don't actually have $500k either, but I'm willing to wait in line because the odds of me finding a project that is really cool and important and could be done with a $500k instrument is about 4 times better than me finding....$2,000,000 USD.

They also rolled out software updates for the whole ZenoTOF line that should be coming close to 1Da SWATH windows with pulsing enabled. I think they were previously getting closer to 3 or 4Da. I really enjoy reprocessing the 2Da DIA window data from the Asstral. Could I get half the coisolation interference for less than half the price? 

Bruker ASMS! No new releases and hopefully a refocusing on support?

 

Edit #3 (6/8/2026): 

This original and then second post generated way too much traffic and too many messages. I've got 2 proposals I'm working on and don't have time for it. So...I took it down and took a little break...

Actual Bruker release at ASMS didn't feature new hardware, which I was very excited about. There is a new diaPASEF variant and some improvements to the PASER/BPS thing. They just rolled out the AIP and TIMSOmi last year and this year was focused on applications of these devices. Cool stuff across the board. Hoping this means a focus on supporting the stuff that is already being sold - that is probably better than we know - because people just haven't had a chance to really pressure test it yet. 

Let's see if I can recapture my original impressions and sentiments without the emotions from what has recently been a bumpy road as a customer of the company. Said rough road from my perspective is that I keep seeing new instruments absolutely everywhere and I have not seen anything approaching a proportional increase in the number of people to support said instruments, at least in my geography. This goes for both field service engineers (the ones that install and fix the stuff) or application specialists (the people who troubleshoot and/or develop new methods). In the latter case I feel like there is actually a decrease and can cite specific examples of emails that now bounce, including to some of the most experienced TIMSTOF support scientists in the country. 

Don't get this mixed up - good, well-trained, qualified people are still out there. But if you go from selling 1 instrument every 3 years in the United States to selling 3 instruments in each decent sized city per year, support should scale. In addition, I've personally seen a rapid escalation in the prices of the instruments and service contracts. Escalation that could and should more than make up for global inflation. It's tough to not be frustrated when you see a company increasing prices, decreasing support and buying every silly looking company in sight. 

First ASMS instrument drop? A new Tribrid?? Check out ApeX!

 

Official site link here! 

Highlights? 75 Hz in a Tribrid? That's rocket fast. Probably rocking the Excedion Pro's lower resolution Orbitrap scan rates? Unclear, but that would be the safe assumption. 

It looks like there will be 4 versions of ApeX aimed at different markets? 

There is a long held tradition for this vendor to take a big 'ol dump on the systems that you currently have.....wow, they rolled out a lot of details on these systems.... but there it is....

...a giant jump forward from the venerable Assend....

If this is the one that is boring enough to release several days before the conference, it should be an eventful week! 

Wait. There is another one hidden in here as well - The Excedion not Pro. Amateur? Excedion Rec League! 

....my brain is off. Obviously it's the 

The upgrade video is a fun watch - the first thing they do is CHANGE THE STICKER! At 1:37 seconds, but then you get to see the differences at the very end. 

PAQu - Integrate transcript and proteomics data to get protein isoform quantification!

 

I find it more helpful these days to simply point out the failure rate of transcript level measurements (because just about every wet bench scientist out there has ran into it), it is relatively cheap and easy to get those transcript measurements. (However, I've still never been offered a $100 genome. Have you? I hear it's a thing, but it still seems like $100s plural). 

What if it still had some value (besides finding point mutations in variant call files, of course!)? 

These authors suggest straight out heresy and suggest implying that you could integrate these data to group those peptide IDs into actual protein data better. Proteoform data thanks to RNA? 

All you need to do is - 

Okay, but since my hiragana is not good nor has it ever been good, I have to just skip over this in every paper. It's just impressive to see this much written out in one block (I didn't even capture it all). AND you can get all the code at this Github. 

What matters is that they demonstrate that peptide level TMT data integrated in this approach improves their analysis. It also lends support to a hypothesis that 2 proteins are very differential in this disease. Two proteins that are so close together in sequence homology that your standard proteomics pipelines probably just lump them into a single group! 

MSstatsResponse - Make sense of Chemoproteomic data!

 

Just leaving this here so I can get back to it later in case we have some drug response data with a lot of variables! 

Super smart and some of the most honest writing. "Yes, in this dataset this other tool actually proves more sensitive"! I love it. The authors use both simulated and real datasets they generated using DIA and TMT and compared them. Refreshing and clever, even if you can't follow the maths. 

Bridging simplicity and depth in single cell proteomics.- some neat observations!

 I do like it when a new group gets on the the single cell proteomics train and starts optimizing/reoptimizing things. Despite the 300 reviews, 50 method optimization papers and 20 biological studies that have been published, each new one brings a new perspectice and observations.

While I don't love every aspect of this paper (some insight on what LC gradients were used when seems to be entirely absent from the main manuscript, which makes me question the title which seems to describe a single workflow) there is some gold in here! 

In my lab we don't reduce and alkylate the single cell derived proteins. I do this because I'm lazy. And also because I spend a lot of time studying drugs that modify cysteines. 

A really nice evaluation of different reduction and alkylation conditions in this study finds optimal conditions and reagents for reducing and alkylating. However, the %CV decreases when doing so under most conditions, probably due to the extra manipulation steps. This section is really well done.

This group also looks at how to digest single cell derived proteins at 37C without those teeny tiny droplets of trypsin evaluating and comes up with a method that works well. They also describe an easy way to get to 100% humidity. 

Since they're using a nanoelute they can go between 96 and 384 well plates. Simply moving from the 96 - 384 well plates is enough to give them 300 protein groups! 

Multiple cell lines were used here. A549 cancer cells, primary astrocytes grown on poly-lysine, etc., it and the system in use is a TIMSTOF SCP and data was processed in DIA-NN 1.8.2 (?) Even if you ran your first single cells (eek. 8...or 9...? years? ago?) There is probably something to learn or re-learn here. I'm certaily adding it to my lab's Slack channel for literature now!

From peaks to power - scans/peak still really truly matters!

 

If you've been on this blog much recently, I am sorry.

Also, you have probably seen me in some level of outrage about some recent studies where people have gotten anywhere from 1-4 measurements of the peptides they are looking at. Is it better than Illumina ProteinCrap? Absolutely. But is it good for mass spectrometry data? No. 

Why is it bad? Because some blogging academic says so? 

This new preprint looks at the problem in depth and finds that for high abundance proteins in blood, the 1-4 measurements per peak is actually not all that bad. Unfortunately....the cancer biomarker you are looking for is probably not albumin, transferring, or immunoglobulins. For low abundance proteins, getting fewer scans per peak means you miss any changes between healthy and cancer patient blood. So....honestly... what's the fucking point of doing the study anyway? 

They say it nicer than this here! 

Nanopores are coming with 150,000 peptide libraries!

There is some replication is flattery quote, right? I forget what it is.

You might need a free account to read this, though. And the stuff from the article that I found most interesting was a link to another GenomeWeb article. Not sure what the rules are for taking screenshots from it.... But the point is that the Oxford people have taken a page out of the ProteomeTools project and have 150,000 peptides multiplexed labeled that they're currently running through nanopores! Smart, right?

Which seems similar to what this group recently published on here, except they aren't working from synthetic peptides, rather LysC digested proteins. 

Nanosplit the transcriptome and proteome from single cells (without the hard part!)

 

When I first saw this I thought - okay, so someone copied the nanosplits paper but they had an Asstral.

And it's almost what this is - 

...but nanosplits requires a technically tough step where you split the droplet containing your mostly lysed single cells. This protocol gets around that step. They still use the same silly robot to isolate the cells, but you absolutely don't need it here (where you basically do need it for nanosplits, it's tough to print that droplet array in a FACs core), and that's a huge win for anyone who doesn't have the slow silly robot. 

Library biases still remain in proteomics hardware particularly for low input TIMSTOF data!

 

I was first going to start with something like this - 

When I read this title 

But I realized that 

1) That's sorta mean.

2) I bet a lot of people thought that all the work that has been done to adjust spectral libraries and deep learning algorithms has been successful

3) Not everyone is doing loads of weird cell types by single cell proteomics on TIMSTOFs and probably doesn't run into this every single day that their TIMSTOF happens to be working.

4) The giant red light on the whole front of my instrument is bumming me out. 

Here is the thing. The Orbitraps had a HUGE head start on data on public repositories. And in the libraries we used to train deep learning algorithms. And every other data type is just different. Especially when you're going down to low load. Even there, we know the Orbitraps struggle against high load libraries. I should put a link in but I can't find it. 

We absolutely find that having reference libraries in single cells helps a lot. On an Ultra2 we like a 25, 50 or even a 100 (for very small cells) cell pool that we run a couple of times and include that in our data analysis workflows. For big studies I've had luck making the library with those 100 cell pools and then just searching the single cells against those new libraries. Now...you'll probably miss that rare cell type and what makes it special, but you might not care about that in every experiment. 

Anyway - this group has some really smart tips for how to build these libraries and the observations in different software. Ultimately they report a 90%(!!!!!) improvement in low load peptide ID rates, so...that's absolutely worth looking at!

S100P levels are linked to recurrence in cholangiocarcinoma

 

It might be easier to make a list of things S100 proteins don't appear involved in at this point.

This paper is going to be posted here because I'm personally interested in it and I wish my lab had access to these samples. 

The samples were digested with some amount of trypsin. You'll never find out how much, but I bet it is fine. They were also labeled with some kind of TMT reagents. The TMT labeled (and, presumably, pooled) samples were analyzed with a Q Exactive of some kind, probably, despite the Agilent high flow coupled Fusion system in the diagram above. The files are on ProteomeXchange if you cared to look. A secret length and flow rate of a gradient of some length you could extract from the .raw files if you wanted, was used for what was most likely a very reasonable DDA method. They couldn't share the resolution of the MS/MS because that might tip you off to what TMT reagents were used. And if they said they used a 1.4Da isolation window someone would complain about it, as would another group if they used a 0.4 Da isolation window. The authors avoid all that controversy by not sharing any of the steps necessary to repeat this analysis of these same tissues.

That being said - the files are publicly available. It could be one of those things where a core ran the samples and the group never paid them, and the core subsequently couldn't find the hours to contribute meaningful corrections to the paper. Also, the downstream analysis seems compelling and it looks like they really thought about their stats in this little cohort. We can probably assume that the mass spec stuff was done right. We can also assume that the reviewers and editors had a lot on their plates when this one slid through peer review. And that happens, we're all busy.