More COVID-19 Protein Resources!

Who is tired of COVID-19?? Me too!

But the resources are awfully disorganized. Here are 3 you might find useful.

1) Newest Protein FASTA!!

Are you directly looking for the newest UniProt FASTA? It will be in 2020_02 (Mid April release) Those great people have already made a pre-release but you have to get it from FTP.

You can directly find it at the FTP site here.

If you get a weird link failure, chances are your browser or your firewall don't allow direct FTP access. Re-enable those if you have the ability to do so.

If you're on Windows, you can sometimes get around that by going to your File Explorer and putting the FTP address in like this:

If you just need the FASTA itself, here is a download link. The UniProt has accessions and is cleaned up and nicer than the one I assembled earlier. I did rename it since theirs just says "coronavirus" (Thanks Virginia Tech for the storage!)

2) 3D Models from SWISS-Model are up!  You can check them out here.

3) That's not nearly enough 3D models for you? Check out the resources from the Zhang lab at Michigan!

These include high complexity models like the assembled units for Chain A, B, C and Spike protein.

Unrelated:

Need Microbiome FASTAs -- This NIH collaboration has you covered!

Thank you to Dr. Nishikant Wase (metabolomics/microbiome expert extraordinaire) for tipping me off to yet another amazing resource.

You can visit the NIH Human Microbiome project here. They have what you're looking for!

I had to bumble around for a bit till I found this page.

Protein FASTAs are "Pep" and they are broken down by microbiome site. How cool is that??

The files are kind of huge, particularly if you're impatient and download all of them at once. (1.1GB and they're .GZ zips, so you'll need 7zip or similar to open them, but who cares? Microbiome Metaproteomics enabled!)

Individual ion protein mass spectrometry clearly explained!

Okay -- around the COVID-19 stuff, one of the biggest things to ever happen in mass spectrometry is possibly currently happening. Like right now.

We're looking at something super crazy that might mean that we've been able to do top down proteomics and MEGAdalton mass spectrometry for -- maybe since some cool Russian guy hanging out in England made an Orbital Ion Trap.

I think Dr. Yates sums it up pretty well here. 3 papers, reeeeeal fast. And several more in preprint or ahead of print right now.

I wrote a seriously inept post about the last one a few days ago. The main point of that post is to remind myself to find someone waaaaay more qualified to teach the Orbitrap (I'm changing it from basic to ULTRAbasic) physics class at/around DC/NIH every year.

I know I've mentioned ELI5 here before. (Caution: This is a Reddit link. You never know what will be there next) ELI5 is:
Explain
Like
I'm
5-years-old

Carl Sagan had a passion about trying to demystifying and de-jargoning science. That dude was writing about it being a problem in his books in the 1970s, and...come on...what did 1970s scientists even know about? Not all that much in comparison. I got my PhD by working out the structure of around 30 glycopeptides. Chris Ashwood knocks out THOUSANDS per run. To be fair, I know him, and he's a lot smarter than me, but science isn't getting less powerful, it is getting far more powerful.

I'm sure that Dr. Sagan would agree that removing the jargon from science now is harder than ever -- and getting harder every second. If you can really truly demystify something, you are a probably a grandmaster of it.

So -- since I realize that I2MS is really super important and I can't figure it out -- I tried to find someone who could ELI5 it.

Introducing Dr. Mike Goodwin -- Ion Physics Grandmaster -- who can explain this revolution in TWEETS.

(I worked with Mike for 2 days right after he got his PhD. Dude is scaaaaaaaaaaaaaaary smart and I often bug him with questions now that I used to only send to that cool Russian guy I mentioned above, who I still haven't met in person, but is very patient with questions from dumb people.)

How I2MS works!

Link to STORI paper above, so I feel like I'm doing something useful here. 

RIGHT!??!?!?!

Okay -- this is probably more like "Explain Like I'm a Mass Spectrometrist" but still!!

Huge shoutout to Dr. Bath, Dr. Ashwood, and whoever is NU Proteomics, some cool people in Wisconsin and a few other people who also helped clarify and focus converations around this topic!!

CoVID-19 / SAR-CoV-2 infected cells real proteomics data!!

Not a proteomics expert and just want to see what happens to protein expression during COVID-19 infection? Here is this amazing web portal that went live this morning! 

Here is a link to the preprint!

Significance statement (if you're not sure why COVID-19 is significant, please go back under your rock) but also -- keep in mind that proteins from these  viruses become detectable in body fluids BEFORE symptoms appear.

2003 SARS Paper showing this is the case in that disease
medRXiV on CoVID-19 that posted yesterday that supports all of this!

Begin standard rambly proteomics blog post:

Remember when the 2003 SARS thing happened and we were useless? With one scan per second and instruments that could just almost detect protein in solution that you could see with the naked eye, mass spectrometry was a medical curiosity. (It didn't stop people from running around saying bold things and building expectations that we couldn't meet, the fallout of which we're all still dealing with now).

NOW --- We can help. It might take some sleep deprivation from a really truly skilled team. And here is proof.

Interested in coronavirus, what it does to the proteome of infected cells, and what different drugs do to it?

I'm not even kidding. They did all this stuff and more resources have gone live since it posted last night.

Here is a link to the RAW files.

Want to interrogate individual proteins -- online? Here is their website!

Edit: And this isn't some crappy rush job. This data is superb. The loading is well controlled, the gradients look great and there is a good balance in the MS2 between the reporter ion liberation and the peptide fragment ions -- I don't have it all downloaded yet, but what I do looks like this ---

-- we don't want to be dealing with a pandemic, but if one was going on these are the kinds of people we want and need out there working on it.

Shameless plug: We've been trying to make peptide/protein detection assays available for as many LCMS systems as we can . We know this resource isn't perfect -- heck, it might not even be good, but we're trying to make resources that are helpful. Another draft is in works with some corrections and updates, but we have to sleep sometime. Here is the first draft preprint.

Use the SCoPE-MS concept to BOOST! Tyrosine Phosphoproteomics!!

They got it to work for Tyrosine phosphoproteomics!! I've worked on this for S/T and it's been -- meh -- and I know another scientist (who actually tipped me off to this idea) who has had some results but it hasn't been a clear winner -- but where we really really need more signal?

Tyrosine phosphoproteomics sucks! The signal is low and they fragment poorly. We can't fix all of this with more signal, but it definitely won't hurt.

Check this out!

As with any of study and, in particular, with anything employing differing levels of loading -- you have to painstakingly optimize everything to get it to work right. And that's what you'll find here!

A synthetic peptide library for crosslinking studies!

What a great idea! There are so many factors in crosslinking studies:
1) Did the reagent get in?
2) Were the ratios of crosslinker to protein appropriate?
3) Did we incubate it long enough?
4) Did we digest the proteins properly given the new factor of this weird crosslinker thing?
5) WHICH OF THESE 10 NEW CROSSLINKING PROGRAMS DO WE USE?
6) WHY DOES EACH CROSSLINKING PROGRAM HAVE 4,000 PROCESSING OPTIONS?!?!?
7) SHOULD I GO BACK TO WORKING AT TOYS-R-US?
8) TOYS-R-US IS CLOSED??? WHEN DID THAT HAPPEN?

The resources and the approach provided in this new study allow us to rule out factors 5-10 (it got increasingly strange for a Monday so I deleted a few):

If you are going to benchmark your crosslinking tool of choice against others out there (they provide some great guidance in the paper by comparing several) you're going to want to go straight to ProteomeXchange/PRIDE 014337 here.

(Minor note, related to a separate paper open on my desktop, please don't forget to cite the repositories and ProteomeXchange in your papers. They need proof we're using them to keep them going!)

You know what may actually be a better use of these Xlinked libraries that contain multiple crosslinkers (DSS, DSSO, DSBU) fragmented in multiple ways (HF-X, ETHCD, Lumos MS3 workflow)?

You can use these to learn how to use the crosslinking programs! There are, for real, lots of options in most of them and now you have expertly created files where you know what is in them!

If you can't find the crosslinked peptides, you're doing something wrong!

If you find too many, including ones that can't possibly exist? You're doing something wrong!

US HUPO Cancelled by some stupid virus thing.

As you might have heard, the awesome US HUPO meeting has been cancelled due to a stupid virus or something. Some people are starting to make what they wanted to present there available.

If you'd like me to help you get your poster or links out, I'm happy to point people to them.

I assume some legit people are thinking about doing stuff.

To legit US HUPO people: If this dumb blog can be an asset in any way, hit me up (orsburn@vt.edu). I'll be happy to post whatever helps here. If you get organized with a big virtual US HUPO 2020. I'll make a big button that says GO TO VIRTUAL US HUPO HERE.

Here are a few of the things that have been posted that I have ran into so far.  As always -- please see my disclaimers stuff over there. ---->

[Just in case: If this is vendor stuff I just post it here because I think it's cool. I never get money or favors or anything for anything posted here, and I'm a jerk to anyone who asks, and can provide references to support this. (Obvious exceptions are in the past when I worked for some companies and when I've selfishly posted stuff my team is working on, which I always feel sorta gross about.)]

Posters:

Biomarkers study from the Goodlett Lab

Great poster on Percolator advances from Will Fondrie and Bill Noble

Multiple posters are available from Biognosys and their collaborators (abstracts up, you'll need to request copies)

Live Talks

PreOn (a proteomics processing robot from PreOmics) will be airing their planned talk live.

I2MS -- We're looking at RAW data the wrong way for top down!

I've had a lot of pictures on my desktop from the I2MS papers. (Paper 1, above and 2, below)

Full honesty -- I have no idea what is happening here. I might try to puzzle it out while I'm writing this.

What I do get -- when I do intact protein mass spec I need a TON of signal and nice chromatography to keep it from giving me crap like this...(from paper #2). In particular, if this is complex? You'll see junk like this all the time.

When I get garbage spectra like this I conclude 1) something big is here and 2) its going to take a lot of work to fix this so I can figure out what the hell it is -- if it's possible at all.

From this same garbage spectra, however, reprocessing this with the I2MS approach (the 2 is supposed to be superscript and I don't know how to do it) -- you can supposedly get something like this --

WHAT?!?!?

Okay -- so you don't have to take the authors' words for this -- you can just go to the supplemental (from the second study) and download the software. It's called CADMIUM. You will need to unzip it twice. And then you'll get a deceptively simple looking interface.

There are demo files in the supplemental. There are only 8 buttons here, but you're going to need to read the directions. If you just start dumping stuff in it and pushing the buttons it'll get angry.

Okay -- full disclosure -- I've been working on this post for like 2 days off and on. Trying to get something from the 2 papers to click in my head so I know what is happening.

I know enough to realize that this is important and I'm comfortable enough with myself and my place in this world to say that I still don't know what is happening. If someone wants to give me an ELI5 or a crayon drawing I'll make time for it!

Single cell (with SCoPE-MS) up to 1,500 Proteins per cell!!

I'm just going to leave this here. 

The most important part of this story is that optimizing really really boring things (AGC and fill times) is critical to getting the best single cell coverage when using SCoPE-MS. This group uses the team iBASIL, but I'm not entirely sure why, except that it is easier to type. I'm not reading super carefully here.

Like the Ecliplse LFQ single cell paper I posted earlier this week, very small columns are used. 30um and 20um in the case of this study.

We also see a really interesting comparison between the coverage that you can get with a Tribrid and a QE for SCoPE-MS when you fine tune things, but it isnt' really a head to head comparison. A lot of factors change. It is more of a how-to for optimizing both.

1,500 proteins? Not a bad place to be!

I guess with single cell proteomics --

That.just.happened.

Reduced proteasome activity in the brains of an aging model!

Okay -- you know when you learn something new that just makes a bunch of pieces all fall in place at once? I owe the authors of this great new preprint for making a lot of disparate data points all finally make sense, but that is a topic for other places.

This is the study.

What is a killfish? I dunno, but they say its a great model for neurological aging!

The main body of this great study doesn't really give you a feel immediately for the massive amount of work that went into this. The Supplemental does, however. There is transcriptomics. And there is TMT proteomics and there is a REALLY nice method for doing DIA proteomics on a Tribrid with varying mass windows. And whoever did the plots is an artist (is Rtist a thing? I'm going to assume that is a thing).

Example?

This is an altogether great example of taking a ton of data (yes. that is 39 size exclusion fractions -- that's how the study starts) and finding the story in it!

And the story is that the aging brain has a reduction in proteasomal efficiency. And if you didn't look at it from multiple directions you might miss it because the proteasome is a complex system that requires a complex approach to fully understand.