New Hats!

 

It'll take me a while to update everything on all the internet things but we can finally wear these new hats openly.

We'll get used to the colors, though orange and maroon will likely always be my favorite combo.

We're moving so I can get started as Assistant Professor in the Department of Pharmacology and Cellular Biology (through the new institute I don't know if I can talk about yet -soon!). I'll also be helping out Stacy Gelhaus as an Associate Director in the phenomenal Health Sciences Mass Spectrometry center. 

I JUST ORDERED SUPER EXPENSIVE BIG HEAVY THINGS and saw pictures of crates with MY name on them. Not someone else's. My name. Weird. Normally "care of" would be expected.  Huge shoutout to my friends at Bruker Daltronic for helping an Assistant Professor's money stretch out enough to cover what I need to start 

Find what proteins are being made RIGHT now with DADPS!

 

This new paper is an improvement over a super cool already existing technology that I had NO IDEA EXISTED AT ALL. 

Did you know that you could dump in something that cells would mistake for normal methionine that you could selectively pull down, so that you knew what proteins were being made at that point in time? I did not. 

I bet the nerds pulling down ribosomes and cutting the attached DNA with nucleases, then busting up the ribosomes and sequencing what didn't get digested (RiboSeq, definitely not convoluted at all) don't either. Sure, RiboSeq is cool - and programs like ProteoFormer have been around for a while combine RiboSeq and proteomics data output. 

But I don't know how to do RiboSeq. I'd have to go back to a grant application that wasn't funded a while back to figure out who on our team was the expert on that part to even know who to start asking dumb questions about the technology. OR I could just do this? And have an output I understand that says "your drug is causing the cell to start making proteins A/B/C right now?" 

Again, this is an optimization, but the authors use the original long acronym thing you don't need to commit to memory (because theirs is better) and demonstrate that is can also be used with TMT, it seems to work best with SPS/MS3 based TMT quan. 

Cognitive impairment at high altitude? Proteomics (and metabolomics) to the rescue!

 

I don't know about you, but I'm waaaaaay dumber than usual when I'm at really high elevations. Not only dumb, but also lazy and tired and 2 glasses of wine and I might just fall right off of a ski lift. 

I went to a wine convention thing in Colorado at a Ski resort years ago and found out all of these things. 

This study tried to get to the bottom of this by collecting proteomic and metabolomic samples from people who went to work in high elevations for 6 months. Holy crap. Some of this work was at 4,800 meters in elevation. There isn't anything in Colorado that tall. The highest lift I fell off of wasn't even 4,000 meters. I'd be useless up there. 

These authors compared serum proteomics (TIMSTOF Pro with....ummm....I'm not sure I understand what else. SDS-PAGE fractionation? Online? At high pH? And an Ultimate 3000 was used on the TIMSTOF but no details seem to be here for the column, flow rate, gradient length) and the data was processed by ....umm....magic....? No software was mentioned that made the PSM/Peptide or Protein Assignments.  I suspect the person doing the actual proteomics was not consulted by the authors on that section. Or they're author 7 and they're like "f' you guys, figure it out yourself" which I've heard sometimes happens. 

The metabolomics was done on a Q Exactive and there are details. Waters BEH 2.1mm x 10 cm column at 350uL/min for a completely secret gradient length. It literally says "over time" and if you want to know about how they ran the Q Exactive - you did not come to the right authors. No joke, check this out. 

We optimized it for best performance and we will never ever tell you what that was and it's weird you'd want to know. The reviewers were not on top of their game for this one. Weird to see something like this in JPR. Nature something or other? Sure. Not JPR or MCP. Meh. 

The data does appear to be publicly available if you're all nosy and want to know or...you know...if you thought this was cool and you wanted to reproduce this work.

The plots are really nice, though. MSStats was used for the super secret proteomics data and an R wizard was onboard who tells you what packages he/she/they used. Hmm...no version info - but AHA there is a github and it is populated. The biology looks cool and I really dig some of the graphs, so Imma post it. The authors could put the complete mass spec methods in the Github later since it doesn't appear to be published to Zenodo or something that will lock it from alterations. 

The single cell proteomics/mass spec meeting schedule for 2025!

 

Image from my kid's very favorite thing to fall asleep looking at - NIH Bioart! 

When Single Cell Mass Spectrometry is announced, I'll add it here. Please ping me if there are others. I have no idea how I missed that Asilomar thing a couple of years ago, but I still feel dumb about not even knowing that I should be there. 

Here are the first two! 

Single Cell Proteomics (US/Boston/NorthEastern) May 27-May 28

European Single Cell Proteomics (I love this one and it crushed me to miss 2024 in one of my all time favorite cities) Vienna! August 26-27! 

SomaScan - 7k vs 11k - seem to largely agree with one another...

 

I'm leaving this here largely for me to check if I can find plasma proteomics by LCMS that I can confidently link back to this same cohort. There is a lot of LCMS proteomics data from these authors, though all I have actively worked with has been muscle biopsies.

However - people are going to want to use the 11k SomaScan assay that is now out there, and it is nice to see that these authors find - after proper normalization  and a lot of batch effect analysis - that data from these two technologies are aligned.

It should be noted that the precision of SomaScan has been shown to be solid. In 8 years of watching for it, I still haven't seen any evidence that the system produces results that are an accurate reflection of the amount of protein present. 

With the growth of this technology - including the use by some of the best groups in the world, such as this one - I hope hope hope hope that it is actually accurate and one day we'll see proof of it. 

HUPO 2025 - worth it for US people for döner kebabs alone!

Planning your conference schedule for 2025 and trying to decide which amazing US east coast city is going to get your money? 

US HUPO in Philadelphia? (February 22-26!) 

ASMS in the greatest city in the world? (June 1-5!) 

International HUPO in Toronto? (November 9-13!) 

If you're already in North America - Toronto is your one chance to get the amazing artery clogging amazingness that every European takes completely for granted. Toronto has döner kebab places EVERYWHERE! Even in NYC - which has everything - you need to really look, and probably hop on one of their famously clean subways to pop over a couple stops to find one. Toronto? All over the place! 

Is this FINALLY PASEF MALDI?????

 

I gotta run, and I'm not going to get my hopes up, but this looks like it might finally finally actually be MALDI-PASEF? 

Get ultralow flow rate nanoLC with ONE pump??

 

I did my PhD with an amazing chromatographer as a co-mentor. As such, it was just easier for him to handle all the hard stuff and I'd just run the vacuum chamber things. I like the simplest HPLC possible. Less to break and even better if I don't have to EVER look at a diagram like whatever the picture above is trying to explain to me. There are definitely switching valves involved. 

However - I could sure think of 4 things I could use ultralow flow rates for right this second. What if I could get that with ONE loading pump by preloading plugs of solvent that naturally form gradients by diffusion? Sounds like magic, but I'm absolutely interested! 

Check this craziness out here! 

Astral vs TIMSTOF Ultra on real single cells? Just about evenly unreal numbers!

Maybe one of the coolest things about the EvoSep is how it really can allow us to minimize variables from one system to the next. You generally have the same column (or, at most, maybe 3 columns, which do make a serious difference, more on that some day), but the same flow rates, etc.,

Which can allow some head to head instrument comparisons that are hard to get otherwise! 

Without further ado - ultra low level samples - 

I don't have either, but I do have ProteomeXchange and 

and 

People who actually make their results publicly available! 

As another front end bit of usefulness both studies used this amazing single cell sample prep front end

Which generates AMAZING SINGLE CELL PROTEOMICS DATA. No question. Amazing. Is it the most expensive way you can prep a single cell today? Also yes. But before we update a preprint with a new clear winner for absolutely most expensive single cell proteomics study ever performed (you can probably guess what mass spec was used...) we'll pull down a crapload of files from these studies and process them with the same workflow.

Thank you DIA-NN 1.9.1 (also, after talking to Vadim I realized I should make separate library free libraries for each study, it does take into account whether you're using .d or .RAW when it makes the library. However, every other setting was left the same. Predicted off the same UniProt human library with all the same settings - and allowing DIA-NN to work out the appropriate windows for mass accuracy, etc., trying to be unbiased. Select file, select correct spectral library. Run. Wait. Again -boom - standardized data processing? 

As an aside - wow - Ultra files take a lot longer to run in DIA-NN. The mass accuracy isn't quite as good which takes longer and then it's got to do the IMS comparisons. I thought the Astral files were straight up crashing because they were done so much faster without that 4th dimension to think about at all. 

40SPD Whisper (nonZoom) for both. 

IonOpticks Aurora 15cm x 75um (which is Ultimate? Worth noting that they've updated their naming protocols for columns recently.) 

You're talking absolutely neck and neck here. Edge goes to the Astral by a tiny amount maybe? I'm getting 3,200 protein groups on the Ultra(1) and about 3,400 on the Astral per cell. HOWEVER, the paper using the Astral is using HeLa cells which has a higher total protein content than the cells the Broad used in this powerful demonstration of the budgetary power these two groups have. I think I processed 6 random cells from the Bruker and maybe 12 from the Astral. Largely because of the time constraints mentioned above. 

Both groups go to 80SPD in the study. Edge appears to go to the Astral at 1,200 protein groups/cell and 850/cell in the Ultra. Again, that might be the larger cell, but it is very difficult to tell. HeLa has this really useful smooth protein distribution, particularly in the high end, compared to most other cells which is why it's such a good cell type for demonstrating a proteomics method. 

Wait. No plots and error bars? Yo, this is a blog post I'm writing while waiting for espresso to do magic stuff to my brain. If I felt smart enough to fire up GraphPad, I'll start doing actual work. I actually did a decent job on this comparison at the time because I thought I was going to buy one of these big heavy things this year. 

Ultimately, these are always flawed comparisons so this is where I'm going to stop. You can go to ProteomeXchange and pull these down and process them yourselves. Both groups get much higher numbers than I do because they both generated actual libraries from their own data. We know that helps A LOT on the TOFs and generally less so with the Orbitrap data because we know what these predictors are based on - but the Astral isn't an Orbitrap and I don't thing we've got a good comparison? Today on what that difference is. 

At the end, though, it's really cool to see that we - as consumer scientists  - we have viable options for instrumentation. We can pick hardware because of how comfortable we are with the software interfaces, or for price or space considerations and then forget about the silly hardware arm's race and start doing biology with these things. 

ALSO - HOW CRAZY IS THIS??? The QC we had at the core I worked at before I went back to fail in academia was 200 NANOGRAMS OF HeLa and 2 HOUR GRADIENTS. I wasn't getting 3,500 proteins on my hardware with that! The fact you can process an actual sample that is 1,000 times lower in concentration and get 3,500 proteins in less than 1 hours is absurdly amazing. I think about this all the time. Where else can you say, yeah, we got 1,000 times more sensitive and faster in like 5 years? Absurd. And all signs seem to indicate we aren't out of this exponential increase in hardware capabilities quite yet. 

Degradator - the protein degradation game!

 

Thanks to Aleksandra Nita-Lazar for sharing this really interesting educational tool!  

Don't want to read? You can play it here. The trick is that your mouse just directs the protein. You can't grab it to move it. 

What DIA algorithms perform best for single amino acid variants?!?

 

Yes! I have also wondered this exact same question. DIA is great for protein level quan but you get lots more peptides/protein, so it works out if they aren't always as high quality as DDA proteomics.

Protein post-translational modifications still don't work as well in my hands with DIA vs DDA. My last check of human samples ran on a TIMSTOF with DDA vs DIA was like 8x more PTMs with DDA. Come on IRB paperwork! I'd love to write this paper someday. 

What about those annoying single amino acid variants that every human being has? Except for the completely normal in every way Craig Venter who loaned so much early DNA that he's probably just the UniProt human database (there are many inaccuracies in this last sentence). 

Don't do it yourself - read this cool open access paper! 

Probably worth mentioning that this is Orbi-Orbi DDA and DIA. An Eclipse and Exploris were both employed at different points. You could probably assume some variation when you go to faster instruments with lower relative resolution/mass accuracy. So...maybe doing this analysis with a TOF might make sense? 

A minor criticism is that the authors did more work than strictly necessary. You could just ignore COSMIC entirely and just go download the XMAn fasta libraries that an amazing professor (who mentored an enthusiastic weirdo a few years ago - crap, maybe it was more than a few years ago) updates every few years. Original paper here. If you wanted single amino acid variants since the 2020 (?) update, I can see doing the extra work yourself, I guess. 

As an aside (surprise -enthusiasm! - maybe it's back for good?)  we struggled a lot at first with my postdoc's CKB type knockout mice when doing DIA proteomics. There are lots of CKs and they are very similar. We'd see CKM(uscle) in places in mice where there is no muscle and the knockout mice were always down-regulated, not knocked out, because one peptide would be attributed to the wrong CK (creatine kinase). That seemed to get a lot better every time we'd get a new DIA-NN or SpectroNaut update. 

What tool would you use for a fasta with a bunch of SAAVs in it? The paper is open access, check it out yourself. Worth noting, DIA-NN's new update specificaly has words about improving proteoform level quan in the new release notes. 

Mass spec twitter might have found a home - has BlueSky finally taken off?

Do you miss the golden days of us sharing all the newest developments in proteomics along with really dumb gifs? 

Twitter was, in my humble opinion, THE best tool for finding the coolest new papers and quick opinions from people who read them.

Then something weird happened to it, and no one knows what. 

Completely unrelated -

We all ran away a while back and have been trying stuff, but get this - BlueSky has GIFs now. AND I'm getting pings every few minutes of some super cool scientist who just joined. 

Maybe there are better solutions than this one, but I can't find anything on Twitter. Half the time I can't even find the website. I swear it might be on a different address entirely. Now, I may not delete my final account quite yet. I've got this growing Excel sheet of companies that I will never ever buy from and companies actively spending money on Twitter is a great heads up for me to check them out and see if they should go on the list. 

I forget how bluesky works. My browser just remembered everything but I appear to be @proteomicsnews.bsky.social

FFPE proteomics AND metabolomics? Here's a protocol!

 

This paper is my first test for how BlueSky performs as a replacement for the once great ScienceTwitter, MassSpec Twitter.

I'm dropping it here with a backdate just so I don't forget about it. We all do FFPE proteomics all the time! Can we do metabolomics as well? I'd thought the formalin and paraffin and time would mess them all up! What a possible opportunity for multi-omics?? 

What's the best software for DDA analysis of host-cell protein contamination in biotherapeutics?

 

This is an interesting analysis for a critical workflow in biopharma/biotherapeutics! 

It's applicable in other places, but where this is used the most is when an antibody drug is manufactured in a hamster ovary cell line or something and you need to make sure that not much is going along for the ride.

In this case, do our algorithms designed for deep global analysis make the cut? In the end it looks like they all do pretty well (Figure 5 is a great summary) though some are clearly better than others. One commercial package I wasn't even aware of somehow (Byos, from ProteinMetrics) actually has a specific workflow for host cell proteins - and you can imagine if they were thinking about that during the design they'd do a good job - and that does seem to be the case. 

Leveraging proteomics to develop an accurate model system for human fallopian tubes!

 

We've eventually got to get away from animal models for human studies. There are clearly dramatic differences between mouse/rat/nematode/yeast biology and human that lead to all sorts of false discoveries. These are so drastic that some funding agencies have rapidly approaching hard deadlines where they just won't fund the stuff.

But human biology is tough to mimic in a bioreactor, even if those things are increasingly inexpensive and easier to use. Even modeling something as relatively "simple" as the blood brain barrier is not at all simple. How the f(ourier) do you model something as important and complex (and amazingly under-studied) as the human fallopian tubes??? 

Like this! 

 Okay - so one way you could do this would be to get some human donor samples and do some really amazing imaging and then dig deep through previously deposited data to help construct yourself a map. (P.S. I love that the authors did such a nice shoutout to my long time neigbors in the JHU Microscopy Core.) Could you do a lot better? What if you also captured physiological function like oocyte transport?!? Could you end up with a map that provided a dynamic understanding of how the system changes during physiological function? That would help as well, but what if you used that information to build the most accurate in vitro system for studying these tissues as you could? This multi-institute team did something like all of this. Cells were grown out to organoids which were coaxed into "assembloids" (we're far outside of my expertise here, so please forgive my interpretations) and by controlling the matrices and how these cells were coaxed do differentiate and assemble, they get there. Proteomics was used along the way (TIMSTOF HT with EvoSep One, diaPASEF, analysis with SpectroNaut) to verify that the system proteome expression profiles.

Even for someone (me) who couldn't follow a lot of the biology/cell differentiation stuff, this is clearly an exciting work. I'd 1,000x prefer having access to engineered systems for my pharmacology work over mouse models. Up that another 1,000x if I those systems were backed up with proteomic data that they were accurately representing human biology.