Holy crap. How did I manage to miss this great meeting every time? The two HUPO's I've attended stand out as two of the best meetings I've ever attended. The 'Murica version of HUPO is just as legit somehow.
In not any sort of attempt at a coherent order while I've got a gap...
1) This meeting has BY FAR the coolest iteration of lightning talks that I've ever seen. Check this out. 1 hour prior to the poster session, every post presenter for the day gets 60 seconds to pitch their poster.
How effective was it? I have notes on almost every single lightning one (except some of the MALDI stuff, I can't write that fast) and I highlighted the best use of 60 seconds, and universally, I couldn't only get to around 50% of the posters that I ranked by the quality of their lightning. People who pitched their poster in an aggressively smart way (hello elevator pitch for the field that Forbes called "the next big opportunity") were swamped. So cool. So pertinent.
2) 15 years from now you'll be at some mass spec nerd meeting and someone will have the 1.74 beers necessary to make a mass spectrometrist silly and they'll try to explain a 3 part joke that JOHN YATES and ILLEANA CRISTAE somehow pulled off in 60 seconds and you'll either wish you had been there, or wish that nerd hadn't started that second beer. I was there. It was the funniest thing I've seen in 2022. I won't try to explain it unless I'm almost at the end of my second beer.
3) Single cell isn't going away. Geez. SO. MUCH. SINGLE CELL. For a field that is obviously this interesting, every company is trying to squeeze in and push out the others. There isn't tens of millions at stake. There's hundreds of millions of dollars at stake, and companies with big money and smart people are making very sly moves to corner this field and lock it down. We're going to need to try hard to not let them. I highlighted something that Ryan Kelly said about single cell "accumulation of small gains." That's the key. This field is still in its infancy and could be something amazing. There is legitimate danger that someone's 3rd fiscal quarter desperation to push a technology could cause a lot of damage. Not sure if that makes sense or not.
4) Match between runs FDR tools are CRITICAL. Unless something has change recently, FragPipe might be the only thing that can do it. I read the paper more than once. I tried explaining it more than once, it took a really pointed question from the ever brilliant Michael Shortreed (who somehow also hadn't sorted it out go Ben's brain go! maybe there is something left) for it all to click.
As we dig for more comprehensive MS1 libraries for match between runs, error propagate like crazy. If you take anything from this largely useless post, please let it be that transferring matches from run to run without any kind of an FDR can be dangerous. Build a really big offline fractionated library for MBR and run some blanks -- you might dig up a decent proteome from your background noise.
5) Fengchao is ONE PERSON. If you are signed up on the FragPipe Github (if you want to learn a lot, you should be, but RIP your inbox) you know who Fengchao is, but you probably suspect that is the name of a team of people. One guy, I met him, and I was almost as cool as when I met the legendary Swedish death metal band IN FLAMES in a bar in Baltimore. Almost. I had trouble listening to 11 of my top 100 albums of all time for over a year because every song reminded me of how cool I was.
6) US HUPO has a LOT of unpublished data flying around! And this will explain, to some degree, the fact there is basically no content to this post. I'm scared to share much I've seen here. I'm pulling up authors from really impressive things I've seen and I'm increasingly shocked by how much I've seen here that isn't even preprinted. Wow.
7) Charleston is super legit. I'd heard it was a really cool town. For you people landing these huge venture capital funds for new proteomics companies, you should maybe stop to think about this inexpensive little town near a bunch of big universities where you might be able to fill your labs with talent that likes it here.
8) BIG PROTEOMICS BY LCMS IS AROUND THE CORNER. People are showing thousands of human samples here and some really cool PI from LA was showing how 1,000 samples weren't enough for her once she broke this cohort down by genetic background, age and gender. She also showed the payoff if you could get to biologically relevant numbers (and humans are diverse, that number is big)
I'm going to slam some lunch and get back into the talks Tuesday afternoon is stacked with coolness. Maybe I can talk about some of it!
No comments:
Post a Comment