Some changes now that RStudio is now Posit, and in ShinyApps.io!

 

I vaguely recall hearing that R Studio or the people behind it had rebranded somehow and thought -- who cares about that nerd stuff? Not this cool guy.

But then I realized that this reorg (presumably they are under some pressure here and there to make money so they can eat and things) affects me somehow, in that I can't just load an unlimited number of free Shiny apps to ShinyApps.io.

AND I can't use the apps that I have there without limit. If you're a lazy person using other people's resources for free in this way as well, you might hit caps. 

Two cool (off topic) things -- what GPUs to get AND calculate the carbon impact of your lab!

As our data gets bigger (and our cohorts, soon, I hope? right everyone? you're doing hundreds of samples if not thousands?) we're probably going to have to start thinking about grown up computer solutions.

GPUs are an attractive solution for many things, but where do you start? 

Check this out!

Also, while digging through this very approachable guide, I found that this author also has a program that can help measure the carbon impact of your lab. 

Sorta funny to have GPUs in the same blog post with considering how to keep from accelerating what is clearly end of days for this fragile little ecosystem we're living in, but.....maybe still worth considering?

iSCMS -- A reason to go to Utah!

 

I know, I know. As someone who gets to commute in on the Thunderdome every day into the world's greatest city, I get this very special ability to make fun of where literally everywhere anyone else lives, but I probably overdo it. Too bad. Get your own blog AND 

Check out the first valid reason to ever go to Utah! 

Explore single cell data (scSeq) with virtual reality!

This is older but I don't want to lose the link and it doesn't fit in any of the open documents on my desktop as I'm running out the door to see what terrors will greet me in the crumbling facility that houses our instruments, so here you go! 

If you've seen my dumb 3D visualization program for single cell -omics data (which I'm still using, I dunno, it works for my brain) you wouldn't be wrong if you suspected this is giving me very bad ideas to consider on my commute.

Ion mobility separation may alter your ideal collision energies!

 

With all this ion mobility going on in everyone's front ends, is it time to take a step back and consider what might be happening to those ions after they head into the mass spectrometer?

This group sure seem to think so! They took a look at what the front end Cyclic Ion Mobility system did to their ideal collision energies, and there are definite effects. It may stand to reason that these effects might transfer to other IMS systems as well? 

Better (and easier) targeted proteomics with DeepMRM!

 

I spent about 2 seconds looking for Gifs with the words "go deeper" in the title before my under caffeinated brain realized that was a poor idea. The best you get is that creepo actor (Leonardo?) in an Inception gif, and it goes downhill from there.

So, here you go. Free software, better, faster MRMs (and PRMs, but it doesn't really go into it) using DeepLearning stuff! 

PepDESC -- the easiest way to actually interpret single cell proteomics data yet?

 

This might be a lifesaver for me right now as I struggle with interpretation of some big SCP datasets! 

A smarter person in a very competitive niche would probably blog about a lot less of the cool things that he/she reads, but we ain't known for smart in Tallmansville WV. We're mostly known for having really extremely long grudges that defy logic....and...mining catastrophes....

PepDesc goes at SCP data significance from the peptide level with really clear logic and through an interface that is surprisingly less than miserable to set up.

I will give you these corrections, however, if you're interested.

The correct installation with devtools is not the one on the Github (on Windows with R 4.2.0) 

It is "devtools::install_github('dionezhang/pepdesc')" 

and "load" has no function, run like normal as 

library(pepDESC)

My only complaint is that your peptide level data needs to be loaded as a *.rda, but otherwise it is incredibly simple.

As an aside, I did hit a maximum matrix in Proteome Discoverer for single cell data. At >1,500 cells I broke the visualization for peptide groups in PD 2.4SP1. I've got PD 3.0 now, so I do have to get some data out the door before I start the full migration to the new software.

 Super impressive work from a solo(!!) author I don't know(?), who appears to have ran single cell label free proteomics on 60 or so cells to support the development of this code(?) Badassery. 

QC tool evolution - from RawMeat, RawBeans and now RawHummus!

 

I stumbled on this one on accident just now looking for a cool Shiny App I use online for generating pretty bar plots for Metabolomics data.

If you've got a Thermo instrument up to the mid-generational Q Exactives, you can probably find an old copy of Raw Meat somewhere. It was from a great little software company that some big corporation purchased and killed all their products. 

RawBeans is the Python GUI app that I use that you can find multiple entries on the blog about. 

If you prefer Shiny for whatever reason, RawHummus is next in line. It was designed for Metabolomics QC, but I don't see why you couldn't use it for any instrument experiment once you've converted your data to a universal format. You can read about it here. Or run an online instance of it here! 

One step lysis to digestion in 1 hour at 70C with new digestion reagents!

 

Wow.....okay.... I did already order this when I saw some charts Ryan Kelly put up at SCP2023 and I've got cells I need to dose on Monday for testing this, but these results are better than I thought they were. 

...and...why wouldn't this work for everything else? Not just single cells? Well...the kit is pretty expensive, but a single genomic library kit is still $500.... If you are using this for more efficient digestion of single or low numbers of cells 1 kit will should go a long way. 

Either way, super cool little tech note! 

$230 million dollars wasted per year making pretty pictures for journals!

 

The best part of this article is that...well....the publisher that published this requires every figure is a vector graphic...?

Which sucks more, Adobe Illustrator or making perfect arrows like this by hand? (From Cleland, 1967) 

PhosphoDisco - Hustle your to phosphoproteomic data clarity!

Full disclosure, I haven't tried the PhosphoDisco yet, but I'm always on the hunt for the elusive package where I can put in phosphoproteomic data and the program will help it make sense. 

Is the PhosphoDisco it? Maybe? The Broad makes tools great enough that they can make sense of SpectrumMill data, so I'm cautiously optimistic and I'll leave the manuscript here.

What you might actually want is this Github link, though, where the documentation is!