Actually, I'm going to start with this awesome Tweet first.
Great thinking points, right?!? This should be hung up somewhere....
A random list of observations:
1) Be sneaky if you have your kids with you and you want to walk around posters. No children are allowed in the poster area. My best guess is that there is a fear they will die of boredom and the San Diego convention center will be held liable.
-I'm no expert on being sneaky, but the poster room has 4 entry ways with doors that only lock from the outside. There is only convention center staff at the single main entrance. Have someone (me, if you see me) walk to a side door and let you and your kids in. Your kid gets a paper cut and that's on you, though.
2) Shimadzu has the best food -- their TOF still only has 100 resolution in high mass mode, but you can't have it all. (I can make this joke -- my work is getting a second one)
3) I learned that despite all of our differences, all mass spectrometrists seem to be unified in their hatred of Budweiser (whatever that is) -- definitely the most common topic of the last 3 hours of the conference.
OH YEAH THE SCIENCE.
Informatics is king! We're continuing to realize our limitations on this end and more tools to improve everything is coming. One thing I can't wait to post -- JudgePRED (best name ever?) isn't out yet -- watch out -- it's cool. Triqler is a much smarter way of doing FDR (integrating peptide and quan FDR) and the python code is all at that link.
Then Bill Noble came on stage and gave one of the scariest talks I've ever seen. I think I can talk about it because he presented it at RECOMB last year and the notes are in Cell here.
So....if you randomly shuffle your decoy database every time you process the same set of global proteomics data --- you...umm...can get terrifyingly different results.
He showed a couple extreme examples just to see if anyone would run out screaming using both large datasets and large and small FASTAs. If he used a FASTA with less than 1,000 protein entries and then reshuffled the decoys (changing NO other parameters) the peptide IDs could differ by 20%(!!!) from experiment to experiment. (This is real footage of what happened next)
Did you know MCP doesn't allow you to use global data searched against FASTA's <1k...? This is why. His team has a solution coming and it might even be available on Crux now. For now -- umm -- let's all ignore it and not let the collaborators know.
I've scanned the other talks and -- I'm seeing a lot of stuff that isn't out there yet.
There was a great PACOM talk that really showed what it can do -- and was delivered with much more confidence that I could have done with Jurgen Cox sitting in the front row. 😅. We all REALLY LOVE MaxQuant and Perseus and the fact we're trying to build things above this lofty bar is a testament to that. PACOM makes some reeeeeaaaallly pretty plots, tables and graphs.
Instrumentation-wise -- I'm a little behind. I got a personal demo on the Q Exactive UHMR last night -- and I totally want one -- but its gonna be a tough sell for me. This is the top down instrument that the field has been needing. Its an Orbitrap that scans up to 80,000 m/z (not m+H+ --- m/z!!) can do 280,000 resolution, has magical front end trapping (essentially making MS3 feasible in a Q Exactive!) and doesn't require the double calibration stuff that the EMR does. Native top down and they have data for mega-dalton top down work. Is it the most powerful top down, intact mass characterization instrument ever? Yeah -- without a doubt. Can I find a biological problem that would mean we absolutely needed something this awesome? Working on it....I'd love ideas....
There is another contender in the DIA software space -- with Scaffold coming in with a really nice interface. I'm really in love with Pinnacle from OptysTech -- and it can do a lot of other things I really really need (untargeted Discovery! -- I.e., great SIEVE replacement), but it's always great to have options. Both are doing live demos here if you're looking at getting into this field.
Coffee down -- time to hit Day 2!
No comments:
Post a Comment