Inductively Coupled Plasma (ICP) -MS for Proteomics stuff!

Hey! What's this thing? This is my INDUCTIVELY COUPLED PLASMA MASS SPEC thing that is being installed on Monday!

What does it do? It uses a frickin' argon plasma beam to convert anything (even metals!) to gases and then basically uses unit resolution to tell the elements apart.

It's pretty sensitive -- this white paper tells you how to set up an older model than mine (!!) to  quantify how many molecules of cisplatin (a chemotherapeutic that has Platinum in it) are in each individual cell by turning each whole cell and into gas phase primary elemental components and measuring the Pt signal).

For those of you who don't live on continents that have banned the ElfSeverer journals -- there was just a special issue about it in one of their journals --

-- I can't access it either. I think the paywall just asked me for more money than Mario Kart for the Nintendo Switch -- and I don't yet have Mario Kart for the Switch.....yes...it's the same game as the Wi U...which I do have...but....

...look how cool that guy is! He can play Mario Kart on a fake airplane (where is the other seat!??!) with his Switch! That could be me, I just have better hair.

Sorry ElfSeverer....I'm going to read some other papers and -- whoa -- this is almost related - check this great new study out!

We use a lot of metal based drugs in the clinic. And we use them cause they kill cells we want dead. However, just like everything else out there, we often don't know the whole story.

Here, this team is interested in the mechanism of action of Plecstatin -- which isn't active until it's metabolized -- making the mechanism of action tougher to work out. New drugs based on ruthenium and osmium are kind of important.

As you might be aware -- Platinum is expensive. You can make a LOT more drug for the same price if you can replace Pt with another transition metal (Google says Pt is 30x more expensive than Ru) -- which is one reason to follow up on new drugs. According to this paper -- when they've done these switches it hasn't been exactly apples to apples. The drugs may be just as effective but the mechanisms may be a little different(!?!?)

They sample at multiple time points and use a load of techniques to figure out what is going on with this drug. All the ICP-MS is in the discussion and it sounds like it was used previously cause you can track the Ru easily with that technique. What they do here is pull-down shotgun analysis with a Q Exactive using biotin labeled Plecstatin-1 and activated drug.

The two drugs pull down very different populations of proteins -- with the prodrug interacting with proteins that just appear to be from initiating a generic stress response -- the activated form, however, is binding strongly to the target (Plectin?)

They do some further stuff where they modify the drug and repeat to try and figure out the important parts of the drug toward the mechanism of action they're interested in -- and how it changes over time and it appears to point to specific bonds as critical!

What is the moral of this story? Ben needs more hobbies? Probably. But maybe -- it's that ICP-MS can be a valuable tool for drug mechanism work and all sorts of other things (mine's primary job will be trying to protect young people from inhaling lead particles...a problem that is increasing rather than what it should be doing) but there is a bunch of other things it can do (and if you have some ideas where 3 minutes of intense ionization with an Argon beam might help you solve a puzzle? reach out! No joke -- it's 3 min per run. I'm worried about how to keep it busy!)

MALDI-TOF + Machine learning = New way of finding drug resistance mechanisms??

How powerful is machine learning getting? For a distraction check out QuickDraw with Google. You draw stuff and a neural network thingy can even guess some of the stuff that I draw! (Those are mine above). It didn't get "beard" or "phone". It occurs to me that I forgot what a phone looks like.

For an example more relevant to the title of this blog -- check out this new preprint!

What did they do with machine learning? They learned biological stuff from bacteria and MALDI-TOF -- stuff about mechanism of action of different drugs on those bacteria!

Am I qualified to review this in any way? Nope. I flipped through the words, saw a lot of math and absorbed none of it. I can tell you they used a Bruker MicroFlex LT and only scanned from 2,000 to 15,000 m/z -- which seems like a smallish mass range for non-digested bacterial cells blasted with a laser -- but appears to provide sufficiently differentiating data. They also used some interesting programs in MatLab that I hadn't heard of for some of the pre-machine learning stuff.

In the end, I think it's a great Saturday afternoon read on something completely different than what I am supposed to be doing!

A new way to think about programming efficiency -- Electricity consumption!!

A pop sci article on this is available here.

The original study is available here.

Here in America this is of no concern to us. We're all trying to fit the biggest supercharger onto the least efficient V8 or V10 engine we can get dropped into the tallest truck we can, so we have 700 ft-lbs of torque (around 900 Nm to you fancy metric using literate types) with which to have flag hauling competitions.

(This really did happen, btw, and I am WAY low on the numbers -- that truck has 30% more torque than I thought....cause....you know....better to have the torque now and not need it and let your  children suffocate on that CO2 build up, rather than need the torque, not have it, and ensure the continued existence of your species.

However -- if you're one of those weirdos who thinks it would be cool if humans didn't just die in large numbers of things that we could prevent now by doing little things like using a little less fossil fuels (weirdos) -- maybe you should stop making fun of C and...wait...I just learned Pascal is still around...(efficient!!) and continue to make fun of thing like Perl and Java (NOT efficient!)

Or...re-engineer the basic architecture to get more efficiency, yo! If PASCAL can do it, I suspect the world can find a way to cut the energy usage of a Java Script in half. (Source: A guy who has compiled maybe 4 things in his whole life and is only awake right now thanks to generic DayQuil (a combination of strong stimulant, liver destroying pain killer and potent dissociative compound for sale over the counter in the US and labeled as a "cold medicine" -- for you fancy metric using literate types)

The Merosome Proteome!

Throughout this blog you'll find whining about malaria proteomics. You may see it forever, until we finally beat this ancient disease. Any time I feel like I'm really getting good at this -- or the next generation of instrumentation is finally going to be the thing to blow this field wide open -- I end up deeply humbled by the experience. The people trudging through this full time will always have my respect.

And -- this new study from a lot of people I know -- and a name or two I hope to know soon -- just opened up an aspect of this frustrating organism that I would have thought impossible -- the MEROSOME Proteome!

Stolen image from this amazing text-book level Nature Review from 2008

The merosome is a stage between the Plasmodium leaving the liver (where they hide in low numbers) and work their way into the blood stream while somehow avoiding all immune responses. I don't think the mechanism is well understood at all, but the impression is that they hijack the membranes of liver cells and use those as their escape vessels.

To be clear -- this is a RIDICULOUSLY SMALL number of parasites. Ridiculously small.

How'd they get enough to work with? By doing an absurd amount of work. They grew one of the mouse model species of Plasmodium in cells in culture. By using a combination of microscopy and PCR -- they captured the merosomes and enriched them.

They normalize the samples by the genome amplification numbers. Ever had so little material that you had to AMPLIFY THE DNA to figure out how many cells were there? Me either -- but what a great idea. Honestly -- I'd guess if you had that little material -- even today's best MS technology wouldn't detect a thing -- and I'd be wrong.

This team uses SCX stage tips to fractionate the material -- and used single shot LC-MS (50cm EasySpray on a Lumos) using a method built for sensitivity (30,000 resolution MS/MS + 150ms max fill time -- not sure if max fill time override "use all available fill" was utilized).

Obviously there is going to be a lot going on here -- cell culture contaminants, human cell proteins, etc., so they included all those in a data analysis -- and scored almost 2,000 protein IDs(!!) from a phase of the malaria disease I wouldn't have believed possible to work with if I hadn't read this today.....

The downstream analysis might be equally impressive as the upstream work -- making this just a marvelous and inspiring study.

ABRF San Antonio Day 1-ish- Sporadic, arbitrary recap!

I LOVE ABRF. I was a fan for years and I'm super psyched that I've gotten to participate in some way or another the last few years.

If you're a core director -- or want to be one -- and you aren't at least checking out ABRF resources you are doing yourself and you institution a disservice, yo!  There are literally hundreds of people around the world who have the same job you do and ABRF is how they compare notes, both through the conference and through materials they put online and the projects that the work together (often completely remotely) on. 

P.S. If you were at the last couple of ASMS things in San Antonio -- My gosh -- It's WAY nicer in March. 70F, rather than 120F. 

And Sue Weintraub was here. And if you can't learn something from 10 seconds of talking to Sue that makes you, your research, and your core better -- maybe you should try another career. Your brain is toaster strudel -- time to go be a politician or administrator or something.

Lets be more positive -- ABRF 2019 SAN ANTONIO SPORADIC RECAP! 

Someone honestly tried to trick me into believing talks started at 8am -- on a Sunday. Then at 9:30 when I walked over from my hotel with some other sciency looking people and --

--- 40,000 high school kids -- looking like they're applying for the circus or as stunt doubles for one of those Disney High School Musical things. (There were a LOT of injured dancers...)  Did I fall for the greatest prank in history? Did these flow cytometrists as well? Turns out the convention center is just REALLY big and there is science on the other side. Favorite slide of the day (that wasn't mine)

iPRG 2019 is going to do MetaProteomics.  HEY Protein Informatics People!  They need help!  Details will be posted here when available.

My talks were stupid, Nothing to learn there.

Brett Phinney did a talk and -- despite the fact he doesn't wear shoes -- he says REALLY smart things.

Do y'all know about BaseCamp?  Brett said he couldn't imagine running a core without it -- it appears to be live updates for teams with a nominal cost. You know how you've got the one nocturnal weirdo in your lab who might fix stuff at 4am? He could just update BaseCamp that the weird ionization stability thing is fixed. So when the normal people come in at 4am or whatever the fuck they do -- they'd know that it was done! Smart!!

Brett had some software recommendations that I'm unfamiliar with, but excited to explore -- namely:

ProStarr
SafeQuant
PDXQC
DIA-NN (the first N stands for NEURAL and the second one stands for NETWORK)

I learned that vDIA is something that I should be able to set up -- I'm just not smart enough. I'm exploring.

I got to see a talk by Birgit Schilling (which I might have misspelled) -- which is a highlight of any conference -- and hear about some recent work they're doing. Y'all should totally check this out!

How good is the wok in the Schilling lab? I'll stop making fun of SWATH for a day -- for real -- because if you're really really smart about it, you can honestly do science with it.

On that note -- what the heck is a SASP? And why do we need an Atlas? I don't know and neither does Google!

And the web address that Dr. Schilling put on her slides is DEFINITELY NSFW. DO NOT GO THERE. YIKES. I'll post a link if I find one.

Moving on -- have you seen EasyPep? I heard from sources I trust implicitly that we should all check it out.  I heard it compared favorably to S-Trap -- and that is some serious praise. S-trap kicks ass. This requires verification.

Day 2: Do y'all know about Jove? New to me! Its a neat mechanism for publishing protocols with peer-reviewed VIDEOS! I'm excited about it.

Honestly -- I have a cool methods thing written up and I was thinking of having it rejected from Nature Protocols. Honestly -- I may just be realistic and have it rejected from Jove instead!

Okay -- I  have a load more to write -- but also this "job" thing to do. Maybe I'll write more later!

Phosphosite-specific Signature Analysis!!!

Okay -- y'all know that depressing conversation, right?

You've prepped a ton of samples

You've phospho-enriched and fractionated and your QA samples are great and your internal QC standards are perfect.

You've crunched the data painstakingly to generate this amazingly well-curated list of proteins and their quan and the relative quan of thousands of phosphosites --- and you're presenting it to the people that procured these priceless samples and you hand over the results.


And...they...quietly...stare...at...you.....
"Wait," they finally say, "Is that it?"
"You're not done, are you?"
"What am I going to do with a list of thousands of differential phosphosites?"

And you think, or if you're me, you actually say,
"No idea! Best of luck!" Remind them that they agreed to the hard stop so you could run to your next meeting!

Cause -- fuck if I know what to do next. Its been almost 10 years since I did my first SILAC phosphoproteomic study -- that OrbiXL only came back with a few hundred confident phosphopeptides. And a few dozen changed. So I looked 'em all up and it kindof made a story. You can't do it that way (well...I can't...I'm getting hella old) with 3,000 sites you got off of today's instruments....

WHAT IF THIS IS FINALLY THE THING THAT CAN MAKE SENSE OF PHOSPHO PATTERNS?!!??!

Check out this GitHub thing. It's full of amazing tools!!

Check out the MCP paper that describes them!

Then write me and tell me if this is finally THE THING WE'VE BEEN WAITING FOR! (Please) You know what I'm talking about. I'm not getting my hopes up. I've fell for it before. But -- for real -- maybe this is it!

Thermo Fisher Cloud now has 1000GB of free storage!!

Okay -- maybe I'm just special -- cause I logged into the TFS Cloud for the first time in a while and it had a popup that thanked me for being a valuable customer and told me my 10GB of Cloud Data Storage was now 1000 GB!! OR 1000000 MB

"More than millions --- thousands!!!"  If you didn't see cold open of SNL last week -- 100% recommended.

This is also known as 1 TerrorByte (TB)

Proof. If you needed a reason to start integrating all your stuff and getting a hang of the TFS Cloud, maybe all this free storage is that reason!

And if it turns out I'm really just special? Sorry for getting your hopes up!

ProtCID -- Hypothesize protein 3D structural interactions!?!?

New on the list of "things I'm super interested in, don't fully understand, wish I had time right this second to read in depth and -- HOLY COW I'm LATE FOR EVERYTHING but THIS IS WORTH POSTING -- list(?)"

ProtCID (which you can read about in this new PrePrint here)

You know alllllll that protein crystal stuff that people have been building up forever and exists in like 4,000 databases of varying quality and Googlability? You know allllllll that protein-protein interaction stuff that is pretty much the same?  What if somebody was crazy enough to try and link it all together so you could access it in one place?

I think this is what this is!!  Is that worth missing 1.28 meetings for? Siri says "ding-ding-ding-ding", which I'm taking as a confirmation, even if that's what she's been saying every 5 minutes for the last hour.

Open Science Framework? A viable option to a 1 year peer review process?

I was just made aware of the OSF recently -- when the best genome I've been able to find for a project we're working on was posted solely through this interesting/crazy mechanism.

You can check this out here.

I may get the details wrong but this is how I understand it.

1) Everything is posted on blockchain. You've heard of it, but probably wondered how the heck it might be useful? They use blockchain to host the drafts AND the reviewer comments.

2) The reviewer comments are openly available to everyone -- just as the results are

3) The reviewers appear to be paid for doing the reviews. "Paid" might mean imaginary internet points or cryptocurrencies that might have value today and not tomorrow. I don't know, but I feel like DASH has been around for quite a while now, though I've never investigated nor invested in that one myself. I learned my lesson with Garlicoin. ($1.14 I may never smell again...)

4) I got a full genome from this blockchain resource. No joke. If you haven't messed with the "next gen" sequencing data stuff you probably don't know how secret this stuff appears to be sometimes. A lot of time you need to know someone to get access. Sure, NCBI hosts the stuff that has been re-analyzed 50 times, but the new stuff? They often aren't open the way proteomics most typically is. Some of it is patient access stuff (pretty easy to figure out who someone is by genetics) but a lot of it is the lack of infrastructure. The Open Science Framework appears to have that infrastructure!

Best practices for MetaProteomics -- Designed by a super team!

In case you were concerned, my obsession with metaproteomics is alive and well. For an update on the challenges and progress in the field that used PSMs to understand whole communities and ecosystems and global warming and who knows what else?!?!?

Check out this brand new review!

Is this everyone in this rapidly growing but still pretty small field? Hard to tell since the "rapidly growing part" and everything...but its more people than I knew were involved.