Draft - Bunny proteomics or how NOT to write an LCMS methods section

Not for posting - maybe keep it to write a tutorial on writing a mass spec method section. 

I should probably start with a HOW TO write your proteomics methods paper's method section. But I'm just all around annoyed with this manuscript as well as whoever peer reviewed it. And the editor, and maybe whatever this journal is. 

I'm also probably annoyed with the premise of injurying bunnies with a balloon. And maybe about the entire state of the world. 

So here we go! 

First, let's fractionate peptides with an Orbitrap Exploris 480. I only briefly had unlimited access prior to the pandemic to a 480 and only have a little data from it. I easily have 800GB of other people's Orbitrap Exploris data (mostly drug data from Tubingen). I was not aware that you could fractionate peptides with one. I'll go out on a limb and say that...you can't.... but that didn't stop these authors! 

(But...for real...the hardware described can not do this. You fractionate peptides on an HPLC, which is probably what these authors did, but we'd generally require the column length, diameter, flow rate, model of HPLC, gradient, column particle size. You'll note we get none of these things here.)

Was the data acquired with DDA or DIA? High flow LC or microflow or nanoflow LC? We'll never know.

Let's get to the results! Oh, cool. They used TMT! So they probably used DDA. But who knows? 

Wait! Let's go back to the methods! 

Or not. Wow. The sample prep is...unique...let's go with innovative. Hey! When there are 50 established ways of doing something you might as well get out molar concentrations of urea and add it and then remove it and add it again, right? I suspect this was not actually how the proteins were prepared based on the descriptions of other procedures, but it might be. Just an aside, TMT doesn't label well in high concentrations of urea. I think you need it below 2M or something. 

Please note above that proteins were considered diffentially expressed if they were 1.2-fold differential. Yeah! Arbitrary hard cutoffs! Is that including unique peptides only? Or are razor peptides included? If unique, how was "uniqueness" determined? There are some absurd defaults out there in the world for the latter which could have drastic consequences for a well annoated canonical protein database. There is no way bunnies have 65,000 genes (probably? I'm no expert) so...the process of selecting uniquenes should be a point of specific focus for this analysis. 

The point of writing your methods is so someone else can reproduce your results. At this point I don't know the starting amount of protein, the intermediate amounts of protein, the HPLC, the mass spec method type, even broadly, the search engine used, the FDR calculations or how the quantification was performed. This silliness should never have gotten off the editor's desk. 

The metabolomics has the exact same problems - with the added fun of a clear fun and strange type for the metabolomics. A google search should have allowed this to be corrected by anyone, at any stage of the submission process. 

Here we do get a gradient! We do not, however, know what the column used was, or the flow rate. Was this DDA untargeted metabolomics? DIA? What resolution? What cycle time (which is a bigger deal in metabolomics than proteomics)? 

Aside here - An S-lens RF of 60 on an HF (no ion funnel) is probably still not ideal for metabolites. You are asking for a lot of in-source decay of these things. Bet you $10 you don't see inosine in these data. Just two very nice hypoxanthine peaks (the former decays to the latter when the RF is too high). 

The good news here is that THERE IS A DATA AVAILABILITY STATEMENT. 

This weekend I wrote several different authors to request data that "is available upon reasonable request", so this is good to see while I wait to find out if my requests count as reasonable. 

Let's break this down, though for reviewers without proteomics experience. And maybe I should just post this part.

This is how I review a paper (or edit it for Proteomes). 

1) Is the data publicly deposited? 

2) Are the methods described adequately to allow me to give this to someone to reproduce in entirety? 

3) Do the methods reflect previously described methods? If they differ from these, are they justified and/or well optimized?

4) Do the results make sense? Are they well described? How many controls/experimental exist?

5) Conclusions - this is where I'm the weakest. I often include in my letter to the editor that I'm insufficiently trained or versed in this biology to interpret these conclusions and I defer to (hopefully) a reviewer who was selected for expertise in INJURING BUNNY RABBITS WITH BALLOONS. Or whatever. 

Yeah, definitely don't post this one either. I've had a rough