Those additional "features" are like the fiberglass body kits you could get for VW Beetles. They magically turned the Beetles into European exotic sports cars OR Rolls Royces. I think you can turn off any driver assist on a Maserati and it will still outperform a Trabant.
See github.com/pwilmart/ASM... for my thoughts. Typically (for DDA), some parameters affect scoring and those should be optimized for each search tool. Other parameters affect search space size (FASTA file, parent tolerance, trypsin cleavage, PTMs) and search space should be the same (or similar).
Is there anything about amino acids in proteins that suggests uniform distributions are at all relevant?
embracing AI and embarrassing AI seen to be synonymous these days.
All you have to do is aggressively smooth the p-values and fold-changes to turn boring Volcano plots into elegant "clutching at pearls" plots (because they look like necklaces)...
Sorry, it was a joke about doing bulk proteomics instead of single cell.
They could always try to average out that noise by combining multiple cells...
The speed of light is now up to 70% greater with our latest software release!!!
The speeds are insane. The men do a 258 km field sprint (sprinting through the fields) and the women do a 143 km sprint. Lucinda Brand still getting 10th after that crash. So much amazing riding today. Chapeau to all the riders and staff.
The women's race looks like it should be amazing, too. They could add excitement by tweaking the announcer voices to sound like they are going over bumpy cobbles during each pave section and spraying them with mud.
Now with white sidewall tires, vinyl roof, and pin striping! Check out these and other major performance improvements at your local dealer!
You have to get paid to be a pro, so, with no funding now, I have to agree.
It is the VC-ome! The sort-of-ome. The proteomics adjacent space. Where your last grant dollars go to die.
Every friggin' day in America is a new low with no end in sight.
What a perfect application for 3-min gradient, ion mobility, DIA de novo sequencing with unrestricted PTMs. Have you thought about cross-linking? Maybe top-down DIA for instant complete proteoform characterization. You could mix the fungi in known ratios and do absolute quant on single cells. Ha-ha.
Well, that explains why so few people live in cities now!
There is no such thing as a post-adolescent male, so keep experiencing wonder. You aren't a dog, so you're never too old to learn new tricks. There are many non-computer solutions to organizational and other problems. Computers can be like the good china/silverware. Use them for special occasions.
Trickle-down proteomics. The rich get richer and zero science trickles down to the science community!
de Novo needs high-quality 2+ peptides with little noise. That's a small fraction of the data. Any/all ML/AI thrown at noise is useless. Deeper characterization (PTMs, splicing, SAAV) can only be done on abundant proteins with *high* sequences coverage. That will be a few hundred proteins, tops.
I don't understand why "A" and "I" are being used as an abbreviation for "bullshit" and/or "piece of shit" when BS and POS were so well established. I guess that is the disruptive part of this new frat-boy-mansplaining investor-defrauding rent-seeking surveillance scam tech.
My feeling is that it is not lack of highly specialized domain knowledge that holds proteomics back. It is lack a very basic general scientific knowledge. Sig figs and error propagation if you measure anything, How to challenge hypotheses/assertions/results if you think your experiment worked, Etc.
(W)right on, dude! The pen is mightier than the Blade GPU!
Does it work properly? No.
Are the people making it acting in good faith? No.
Have they ensured it can't be used maliciously? No.
Does it make people's lives better? No.
Will it make people's lives better? No.
Is it profitable? No.
Is it popular? No.
Is it harmless? No.
The argument for AI.
Silicon Valley turns venture capital into Billionaires. They seem to make plenty of money!
If you use Microsoft AI, the agentic workflow is just Excel macros written in Visual Basic. That is the literal definition of "stepping in it".
When it is in the crate, it is Schrodinger's MS. It exits in a state of both new and unused and used and un-new...
It's like a different cow pasture with orders of magnitude more cow pies! Watch where you take that first step (and the next, etc.).
Apologies. I had not had my coffee yet. The joke was rather stiff and lacked depth...
Part of the Ben-there-done-that series?
