Your last day to submit a workshop idea for US HUPO 2024! Abstracts open!

 

It's time to submit abstracts for US HUPO!!! 

AND TODAY IS THE LAST DAY TO PUT TOGETHER A WORKSHOP! 

As proven by Ben2 and Brett's awful LFQ Battle Royale Workshop, they'll let anyone put them together. I really just applied for a workshop because I had these hilarious old pictures of my workshop cohosts and i wanted to submit them somewhere and - BOOM - we had a standing room only workshop!

Reminder -- in 2 weeks we'll do round #2 of the Battle Royale open to anyone in the world to see how global proteomics quantification methods stack up. Register here! 

BlueSky Social Week #1 -- Twitter without gifs -- and ads - now BioRXIV support!

 Article here! 

First question -- BlueSky is by invitation only and, no, I can't invite you yet. I haven't been on the platform long enough to invite anyone, but if you were looking for #MassSpecTwitter and for the coolest papers, Mastodon and BlueSky have them.

I need to figure out how to get into Mastodon again and see how it has evolved, but I might already be following more scientists on BlueSky than I ever did on Twitter. 

I did put one post on Twitter this week so I could amplify a student's 63 page paper (!!) he submitted and posted on Biorxiv since Altmetric tracks Tweets but does not yet track reposts on BlueSky, but with bioRxiv joining the party, it's only a matter of time. 

Question everything? Sensitivity tailored DIA for single cell proteomics!

Woooo! Okay - so there have been some surprises in working with actual human cells that everyone is running into as their getting going. A lot of the things you know will work from your years of doing proteomics with basically unlimited material....don't seem to work quite as well. 

And this great new paper goes through a lot of these surprises.

For bulk proteomics we've been trying to get to faster and faster instruments so we can get to smaller and smaller isolation windows, particularly for DIA. We also know from - forever - that longer gradients basically lead to a linear increase (not a 1:1 but I think it's still a line, there probably is a word for it, but I've got a bunch of deadlines) in peptide identifications. 

Check out the image at the top of the post - doubling their gradient gets them what? a 10% increase? You doubling your acquisition time for 10% more peptides? Me either. You can look at the window optimization inside, it's cool stuff. Also -- gas phase fractionation (where you make libraries on an system by doing DDA on little narrow windows of your samples, for example only doing DDA [or, I guess, DIA] from 400-600 m/z, then running that sample again and only doing 600-800 m/z, etc.,) has a lot of value. 

The coolest part here is the biology as they track the proteome of this cellular population as their differentiating. We're getting out of the methods development side right now and it's time to start rewriting some text books (or whatever they use now). 

Make beautiful figures out of any spectrum from any instrument with this easy online toolkit!

 

I didn't want to post anything on this super cool new online tool from Goodlett lab until I had all my figures made. Don't want to Orsburn Effect something that I need for a deadline and talk next week! 

Do you just want to take a mass spectrum from anything - maybe even different instruments or online data or NIST fragments or literally anything else and just make a nice figure? 

www.mass-spectrum.com.

Feel free to read the nice tutorial - or just use the most intuitive interface you've ever seen. 

YOU CAN EVEN MAKE FIGURES FROM BRUKER DATA! No joke.

That figure at the very top is me extracting diaPASEF data from a drug standard and from a drug treated complex sample! Bottom is my drug standard. 

How do you do this? You get into the hellscape of terribleness that is "Data Analysis(TM)" and you find your spectrum and make it a "Compound Spectrum" then you right click on that spectrum in your table and export it as a peak list .ASC file. Then you open that in Excel and trim out the 3rd column that is just random weirdness, save it as a .text file and load it into www.mass-spectrum.com.

You can add an Excel file with your masses/peak lists and - BOOOM - you've made a labeled plot. Everything from the colors to the fonts to - you name it - is user selectable. 

I've successfully made a mirror plot from QE PRM data for a standard small molecule and from a TIMSTOF diaPASEF small molecule run! 

Oh - and there is a paper!

EvoSep Webinar 9/28 - Is this the OneTip unveil? Digest, clean and load on 1 EvoTip?

 

You've probably heard rumors (and maybe there is now a preprint, I can't keep up) that there is a method for directly dropping your cells onto an EvoTip and then digesting them there, cleaning up the digest and then just directly shooting into your instrument? 

Don't quote me, but I suspect this is where we'll see it. 

You can register for it here!

InstaNovo - Can neural networks make real de novo peptide sequencing a reality?

 

Full disclaimer - I can't follow all the words in this new manuscript. It is very computer science term (?) heavy. Honestly, if I hadn't found on page 35 that this code is available it wouldn't have made it on the blog, but from the proteomics data I can follow it looks really promising.

If you're a computational nerd person, I think this is what you want (Github). 

From what I can get, at very reasonable FDR, InstaNovo is identifying as much as 50% of human peptides that are known - with no database at all. None. Sure, having a database for something you have one for looks better, but this opens up a tremendous number of things that we don't have sequences for at all. They pressure test this with less used enzymes (GluC) and do some HLA/MHC peptides and some mixed proteomic samples (metaproteomics). 

"Alternative" careers after your degree! -- US HUPO Webinar tomorrow 9/26 at 1pm EDT!

HEY! You're getting a degree in some academic lab! Great for you! I bet everyone around you where you're putting in a very rational number of hours to get that degree has insight on how to best be an academic.

Does it feel like  

? Probably does, right?

Of the people I started my PhD program with, there are very very few people in academia right now, and I went and did random stuff for like 10 years before coming back. There are other paths. Most of them will be other.

US HUPO has a thing tomorrow talking about 2 you probably never knew existed (I sure didn't).

You can register here! 

Ummm...single cell lipids at analytical flow rates?

 

I'm....very very surprised by this one, y'all. Legitimately I feel like I'm off an order of magnitude or two, but this study seems well thought out and the results look okay on the surface. 

And this looks just....sophisticated enough to be something people would do single cell analysis with

This group used a standard QE Plus operating at 300 microliters per minute (or 350 microliters/minute if using a special column) -- holy crap. I wouldn't have guessed I'd have detected lipids in 1,000 cells. Let alone one at a time. I've got a QE Plus! I'm pretty sure the LC on it can go that high, I generally run it lower so I can have higher sensitivity, though...

Now....apparently the reviewers at ACS did forget to make sure that the RAW files were publicly available, but maybe they don't require it because who cares about lipids? 

Okay, because I have a rapidly approaching deadline -- what is the lipid content of a single cell anyway? Table 7.1 in this old book (open access here). 

5% of the Eukaryote cell content is lipid (probably put a big -ish on this, I can't find the original study of Alberts, 1983 through the references). So....let's go with that's around 50 picograms of lipids per Eukaryote cell. Okay, at least it is more than the 1pg or so of mRNA floating around (about 10% of RNA is mRNA, most of the RNA around is the stable stuff, 16S, etc.,) I'm getting that number off of it being about 1/4 the percent content of protein and a mammalian cell is generally around 200pg (blog post from 2014....ouch...wait, I was still using SpongeBob memes in 2014?) 

What about their standards? They used the ones we use from EquiSplash and they optimized at 16ng/mL

So, 16 picograms/uL and they put about 5uL injections, so they optimized their method around 80 picograms.

That's not too far off the 50 picograms per single cell, except that's total for phospholipids and other lipids. And since I have no idea what the distribution of different lipid species are (how many and do classes 1-7 make up 90% of them) but -- this actually seems like it is in the right realm order of magnitude or two. Weird, right? 

Single bacterium proteomics!

Wooooohoooooo! Okay, so I mentioned that this was coming during my barrage for ESCP in Vienna, but it's already out! Told you, JASMS is on fire right now.

This is a SCoPE-MS style study with single E.coli. Yes, something like 1% the size of a human cell. 

You should check it out. My kid has a potato and he won't stop screaming "I HAVE A POTATO" until I acknowledge this fact to his satisfaction. Really exciting stuff in both the paper and this ready-to-microwave Russet. 

Ben Katz analyzed WD-40 and it is scary stuff.

 This is off topic, but I was looking for a video where some buffer components were dissolving to a really intense dubstep drop and found this TikTok where Ben Katz did a headspace analysis of WD-40.

And....he uses the hashtag #PFAS for very good reasons. If you aren't familiar with WD-40 you probably haven't ever done your own car repairs. I'm a big fan, spray it on a bolt, come back in a few hours, break that bolt free -- POISON YOURSELF AND EVERYTHING AROUND YOUR HOME WITH FLUORINATED HYDROCARBONS. 

This could use more attention. 

Hate your nanospray source? 3D print a better one!

 

WOOOOHOOOO!

Okay, JASMS (Journal of the Ablebodied Sapiens of Mass Spectrometry) is on FIRE this month. I have to really read the next one before I post on it, but I don't have to take notes on a print out of this one to figure out that I like it. 

The pages are labeled A, B, C, D and D is just references! 

Just a reminder you've always been able to make your own  nanospray source for Thermo instruments thanks to UWPR (last I checked in on it, the source would cost you about $150 to build. Given inflation the last 4 years, maybe $350 today?) 

But I ain't seen one for anything but Thermo instruments and this is a really cool step! 

It's September 21st! There is a mass spec symposium all day in Rockville today!

 

I'm surprised that whether I go through Mount Airy or not will only add 1 minute to my trip. It ain't exactly a pedestrian friendly place. Either way, I've got to get going for some free parking! I want to see this charge variant analysis thing applied to antibody drug conjugates! 

Is it time for Proteomics Social Media to head to the Blue Sky?

 

Hey - I tell you what, I LOVED Twitter. It said I had an account for like 11 years? But I really didn't start using until a lot later than that. I bet 75% of the papers that have ended up posted on this blog were from Twitter. 

Then - you know what happened?  - I don't. Reddit uses terms like "Phony Stark" and "Apartheid Clyde" and puts pictures of Data from Star Trek TNG beating up random aliens. I don't know what is going on or what any of that means. I was a huge Star Trek TNG fan (having 2 channels in the 1980s/1990s will make you a fan of anything) and I wouldn't fight Data either. But now Twitter is called XXX or something and there are ads everywhere.

I tried to abandon ship when it first started getting weird but the content coming in was so good and convenient, but now it is just ads. And ads, and randomly a reminder that basically every person in proteomics is an old white guy (something I really truly care about fixing the perception of, because, while it is mostly the case, it isn't completely the case) 

Smart cool young people have been talking about BlueSky (yes, Jordan counts as cool in my book. Probably intentional wacky probably ironic 80s look? Check! Cooler if he wasn't being ironic about it? You betcha! Cool mentors?  Paul Stemmer and now works for Birgit Schilling? Crap. That's some pedigree. Check! Is trying to do wacko hard PTMs by DIA? Fucking right, that's a check in the cool column.) for a while. (Just throwing this in because Jordan was waaaay ahead of the curve on this and all the genomics bioinformatics people have already jumped ship and i feel like an old nerd.) If you aren't on the inside "Proteomics Radio Hour" group of people, I'm talking about this rising star in our field. 

From the picture at the top- it looks like even a slow person like me can pick it up? That looks the same? Do I not have to learn anything at all? Hashtags don't work? I didn't know how to use those anyway! 

And this weird guy I've heard of put up resources! Check them out here! 

Honestly, maybe Mastodon is still the answer, I don't know, but I feel like my publicly (and self) funded research is lending credibility to something very bad now, so it's time to go. 

Improve every pull-down with chemical acetylation of your ligands!

Do you pull-down, affinity enrich, and/or immunoprecipitate things? Want to reduce your background so you can actually see what you're looking for - in every experiment? Do you simultaneously want to feel dumb for not thinking of this first? 

This is easily one of the smartest things I've read all year and I bet you'll agree! 

It's a wrap! THE Proteomics Show Podcast Season 2 ends(?) with a famous science author!

 

Do you know Dr. Rachel Lance? If not, she's a super cool scientist and science writer. She writes for Time Magazine sometimes, and Wired a lot, and wrote a best seller about historic submarines and she gets people to go into pools and then throws bombs in the pool with them and get this -- she's interested enough in Proteomics that she agreed to be on THE Proteomics Show Podcast. 

Are you asking yourself....wait....didn't you say something about the US Human Proteomics Organization sponsored your show for a second season so you could interview the people who make US HUPO what it is? Yes, they did, but you know how the plenary at a big conference is always someone loosely associated that does really cool stuff and you wonder "did the people in charge just do this because they wanted to talk to someone cool and famous (but maybe they didn't want to offend anyone by bringing in a plenary from inside their own field)?" We had some people who couldn't do the show so we interviewed a famous scientist and writer. 

Don't like it, get you own proteomics podcast. I'll totally listen to yours! (Not joking) 

Proteomics reveals that METH (yes, real METH) is bad for mice! And why!

 

Methylamphetamine isn't a funny thing. Writing it METH like you're screaming it 212 times in a proteomics paper, after giving it to a bunch of mice, also isn't funny.

Then why do it? Because as bad as METH is, we don't actually really know why it has many of the really terrible side effects we know it does. 

Proteomics speeding to the rescue! 

Not funny AND that's a rabbit, I think.

In this study these researchers got a bunch of meth and put it in water bottles and let the mice drink it, or have the option to not drink it, or drink a LOT of it) and they did the last thing.

Then they let the mice recover from being METH addicts and they did the logical thing in every regard of chopping them up into little bits and doing SWATH based proteomics on them. 

Don't worry, this isn't just one of those gave mice drugs and chopped them up studies, this group found a lot of new information on the fundamental causes of METH toxicity. AND they found a really pivotal role that Melatonin plays in recovering from METH addiction. It's an altogether solid study and worth a read. 

The Quantum SI is out! Slowly sequence one protein at all! No mass spec required!

 

Do you just want to sequence one protein? 

Do you really want to avoid those weirdos at a mass spec core? 

Do you prefer instruments you can carry around if you want? 

Well do I ever have a solution for you! The Quantum SI Palladium is finally out! They hired a friend of mine a long time ago and it was all secret what she was doing.

You get a PCR sized thing and you never go to that mass spec core again as long as you only want to do one protein at a time. This thing sequences your one protein on your benchtop #1 or number 3 or at lunch at a restaurant with you. You buy these reagent kits and while it is error prone, you give it so much protein that it eventually gets it right! 

Actually there is a paper in Science about it, but it's paywalled. 

DISCLAIMERS OVER THERE! I literally don't know how this thing works. 

Curtain PTM! Visual analysis of your PTMs against libraries!

 https://twitter.com/i/status/1702634612579578292

Want to dig through some protein post-translational modifications in a visual format? Hell yeah you do! 

I feel like Curtain is in review somewhere and a whole bunch of stuff was dropped just now so they can talk about it at the Human Proteomics Organization World Congress in Korea, where I will not be.

There doesn't appear to be a paper, so you don't have to read anything! Just start pushing buttons here!! 

https://curtain.proteo.info/#/

I approved like 200 comments on this blog!

 

What are the stages of grant writing? I'm somewhere between writing multiple book chapters I was procrastinating on and regretting every decision I made regarding my career since 2017.

So...I can't disable comments on the blog for some reason. Probably because I'm dumb, but I put up a post around 2020 that said "please don't post comments" because I get about 14 things telling me how to buy narcotics with Bitcoin each day and a whole lot more telling me how I "can get my hair back" and I can't imagine how the latter could possibly apply to me. And I don't support purchasing narcotics that aren't prescribed by a real doctor, and if you were going to get them, 100% absolutely traceable Bitcoins - because that's the whole point of the blockchain that fossil fuel eating stupid thing is based on - is probably the worst way to purchase them. 

In there somewhere are people who have put really smart comments on these not nearly as smart blog posts. I found like 200 of them! If you get a weird notification from me from 3 years ago, this is the reason.

What a difference 3 years makes in single cell proteomics!

 

Has this been on here already? I forget. If not, you should check it out. 

There was this sortof secret race going on behind the scenes the last 8 months or so as at least 3 different groups were all trying to publish wide window acquisition for single cells at the same time. Maybe there were more, but I was only a top secret peer reviewer for 3 of them and they all concluded that single cell doesn't work the same way as unlimited protein material in that wider windows really are better, but that's not the moral of this story.

What I want to talk about is what a jump this is in such a short period of time from this one -- 

it says 2021 but these fancy monarchy people take forever to construct a PDF. This was accepted in June or 2020, so it's been right at 3 years. 

I've got a paper due to an editor like yesterday or something and as part of said paper, I've been really looking at these data. 

Check out the copy number distribution shift. Sure - they got more proteins per HeLa cell with 3 years of improvements - and - I guess this might seem silly and obvious, but the shift toward the lower copy number proteins is even more striking to me, because most of the cool proteins can be found on the left side of the plot. 

This was plotted with my Shiny App I made a couple years ago, btw. It took me a month to make overlapping histograms work right in R and I'm going to use this skill like Naruto with the whole shadow clone thing (if you aren't familiar, it took him forever to figure out a simple trick everyone else easily mastered and he basically just used it for everything for a comic book that ran for at least 10 years? I was very surprised to see costumes for characters from this ancient comic book in a local halloween store a few days ago, so I guess it's still around? 

Even more impressive (about the progress that is supposedly the topic of this post? The number of unique peptides per protein went from 4.6 to 8.4! Using basically the same hardware and sample prep. Sure, some tweaks but this is mostly instrument method and informatics. 

Super impressive progress in the left direction! 

The overdue European Single Cell Proteomics 2023 meeting post!

 

It was 100F in Vienna and the great people at the IMP put on one of the greatest meetings I've ever been to. I can't do it justice, because this meeting was just too good. How good? There were ice cream breaks and while we're still screwing around with "hey! I can see proteins in a single cell" on this side of the Atlantic, in Europe they're actually applying it to help people. For real. IMP is working with real surgeons. Pharma people were everywhere at hte meeting as well. Application, application, application! 

Right before I got there the IMP received a giant check to build up their capabilities! Ever seen a giant check for science? I sure haven't! 

Also -- ESCP sold out early at 200 people. It was packed. 

Quick run down of day 1: 

Karl Mechtler himself did a stage appearance and said a few words before handing it over to EuPA's leadership (currently in changing of the guard status) and it kicked off. 

I've got to be really careful with photos and what I talk about - there was A LOT of unpublished data. I got all inspired and stayed up most of the night to change my slides and show what I was currently working on. Which...now I've definitely got to write up...but it was that sort of vibe.

If I miss anyone, it isn't a slight, I just have a real job and this isn't it. 

Bogdan Budnik showed crazy new data on human neurons and drugs. He also demonstrated some really really cutting edge software his team has developed and have in works. He also very succintly walked us through the streamlined workflow his group at Harvard/Wyss uses to complete around 400 cells/day using a Q Exactive (I'm not sure which type) and an Exploris 240. Stunning data reinforcing that single cell proteomics isn't all about the hardware. Extreme paper watch alert on this one. 

Lennart Martens gave basically a what's what on CompOmics -- tools I knew about and ones I didn't. If you haven't been here in a while, it's a good time to visit it.  https://www.compomics.com/

James Fulcher walked us through Nano Droplet Splitting. Which is, btw, the thing that gets all the fancy genomics people excited about single cell proteomics. When I say - well....you could still get your largely useless RNA out of each of these cells and I'll still have material to do the stuff that is phenotypically relevant, they get excited. As long as they get their largely useless but exceptionally pretty pictures out of a population of cells, you've got an ally. (Wow. That sounded sorta mean. I like pretty pictures as much as anyone.) 

(Fulcher, 99% sure everything on this slide is covered in his preprint) 

Alexander Ivanov showed single cell top down. A lot of that has been featured on the blog and on the podcast, but I invite you to check it out.

ESCP promoted a lot of really exceptional students. Abel Vertesy, Anuar Makhmut and Suniya Khatun all showed currently unpublished data that I don't want to discuss until I see it preprinted. All reinforced how much better the next generation of scientists may be at proteomics than us. 

I worked on my slides at lunch and ran over a little and missed Anjali Seth and part of Ryan Kelly's talk. We have spoke at a lot of things together over the last year or two and I feel like I'm mostly up to date. 

Valdemaras Petrosius showed off the Schoof lab's fancy single cell statistics and some comparisons between Orbitrap Eclipse and Astral. Most of that has been featured on the blog somewhere. I got to speak to him after and he's been doing mass spectrometry for about 2 years or something. Scary good data 2 years in? Didn't ABRF conclude that it took about 5 for us to generate solid data? Curve appears to be steeper some places than others. 

Day 1 wrapped up with Peter Pichler giving a history of surprising history of medicine/surgery with dates so far back in time we only have archaeological evidence. Then we did a cookout on the roof of the IMP! 

Day 2! (Sabrina Richter on why we need proteomics when we've got single cell RNASeq and all their sophisticated informatics) 

Day 2 started strong with Chris Rose (Genentech) and then Erwin Schoof who I finally got to meet in person and one of the great student talks I mentioned above and Matzinger Manuel (who has the most engaging illustration style for figures I might have ever seen -- I haven't been able to catch up with him to see if he literally draws them or if it is some AI graphics package).  All of them have had work featured on this silly blog recently, and maybe I'll put in some links if I get a chance. Off to preschool in a second.

(Found a pic. How cool and engaging is that?) 

Vegvari Akos (also finally got to meet, terrible taste in music, but really cool guy) went after a question that I think about a lot. Could we do less sophisticated systems? Like -- since bacteria have proteomes mabye 1% as complex as ours, could you gain enough in your reduced "noise" level to quantify things? Again, a lot unpublished, but I think it is fair to say that the number of IDs wouldn't wow anyone, but there is potential there.

Sabrina Richter (pictured above) demonstrated some super sophisticated statistics applied to single cell proteomics. That lab is just killing it. 

We had a minor schedule alteration where we upgraded from Matthias Mann to Dr. Florian Rosenberger for deep visual proteomics. As always, intimidating stuff out of the Mann lab. I was a little disappointed by the resolution of the first deep visual proteomics study but the resolution they have now is absurd. Ugh. Like I really want ot type a word here and I don't think it's preprinted yet so I shouldn't. It has 4 syllables, but I can't type it. Crazy data coming.

Putting here for proof people invite me to give lectures (having some problems getting my expense reports approved). 

My grandmother didn't make it through the pandemic and I miss her dearly, but I'm glad on some level she didn't live long enough to see this picture of me being seeing in public wearing 15 year old jeans. And...to be honest...maybe I need to go shopping for clothes more than once every decade....might need to work that in somewhere. 

Then the conference got CRAZY. Not because of the ice cream break.  Francis Berger's talk was amazing and I don't know how much I can talk about it.  I've never wanted to see a live video of a brain tumor being removed. I've never wondered how many cells you can take out of a certain area in a human brain without killing someone, but it isn't many. And I've never thought about a zombie cell. 

The day wrapped up with Michale Caldwell (NorthWestern) talking about the work some Kelleher guy showed at SCP in Boston I mentioned recently and Christoph Krisp (Bruker) showing what you can do at 300Hz with single cell loads (300 f'ing Hz!) and a really cool breakdown from a Nature Editor about their new directions. They're being super supportive of next generation proteomics and it was cool to hear more about their directions.

The conference wrapped up for me with dinner with some really cool new friends (and Jarrod who is the least sketchy Australian you'll ever run into). 

There was a great tour of Vienna the next morning, but I was on the way to the airport so I only got to yell funny sounds from a taxi on my way out. 

Okay, so maybe I'll fill in some links later to different papers, but if you took anything from this rambling today it should be that EXTREME PAPER WATCHES SHOULD BE IN EFFECT. For real, I've said before that single cell proteomics is mass spectrometries moonshot. I'll go one further

Single cell proteomics is the Moonshot project for biology and medical research. What is going to come out of it as it continues to become less "cool trick" and more application is going to change things. The bigger impact may still be what trickles down from it as people start thinking "10 cells worth of material? That's a ton!" 

HLA peptide identification in 200uL of plasma!!

 

https://www.biorxiv.org/content/10.1101/2023.09.05.556309v1

HUGE ramifications for our understanding of host-pathogen response reutilizing those systems for good. Crazy paper, I pulled off the road while driving to post this. Looks like better grammar than usual! 

Oktoberfest -- Easy local PROSIT deep learning/rescoring stuff!

IT'S SEPTEMBER!?!??! Okay.....deep calming breaths..... ready for Oktoberfest? No, not a crappy 2000s coming of age comedy that is probably reran on Fox networks during the day for the 11 people in the world who can watch 19 minutes of commercials for every 2 minutes of crappy 2000s coming of age movies.

THIS OKTOBERFEST! 

 

Honestly, I'm just typing really fast while waiting for my espresso. I haven't tried this, but I probably will. Easier deep learning tools? Sign me up. 

They do test this versus earlier versions of Prosit and with what appears to be HLA class 1 and it is much better and it has functions that you can only access on PrositDB while waiting your turn. All good stuff!