If you find the news headlines don't make any sense, add the two words "for Billionaires" to the end of any headline. Even if you still can't figure it out, at least it will be true.
I never bought the concept behind the nano particle halo crap. It looks like an excellent way to launder money, though. They can open up bead shops in Haight Ashbury in SF and on Telegraph Avenue in Berkeley to liquidate inventory. Shop lifters with magnets could be an issue...
After you are done with the vibe biology, try vibe unemployment, vibe pan handling, and vibe homelessness! I have got to run now to work on my vibe depression...
'OpenAI, the company behind ChatGPT, is on track to lose $14bn this year. Its market share is collapsing, and its own CEO, Sam Altman, has admitted it “screwed up” an element of the product. All it takes to accelerate that decline is 10 seconds of your time.'
www.theguardian.com/commentisfre...
Does anyone really want an online journal club for all papers? That would be reviewer 2 all day, every day. It would teach the important life lesson that any step forward is like walking through quicksand into a 200 mph headwind with one leg tied behind your back...
Those are the strongest candidate peptides because the PTM will shift all of the y ions. That makes the identifications and site localizations 100% certain. Ha-ha.
Label free is more like 20-30% CVs (despite claims). Protein quant from tryptic digests is *always* just weighted spectral counting. Protein quant doesn't have to be accurate/precise (that helps) but it needs to be comparable/interpretable between samples.
15 is 15. It's counting statistics (Poisson), so std dev of 15 is 3.9 (square root). If you use plus/minus 2*SD for 95% confidence interval, 15 could be anything between 7 and 23. 100 ions is anything between 80 and 120. You won't get decent quant with small ion numbers no matter what analyzer.
Do the proteins from the "fast" cells get to the TOF detector first? Single cell racing could be the next big (or very small) thing!
Well, it is a male model of "intelligence" built by "males" and is almost a proof by exhaustion that there is no such thing as a post-adolescent male...
In the 35 years since Python was first released, lots has changed. One constant has been that "Learning Python" by MarK Lutz is the best book out there. There's a 6th edition (learning-python.com/about-lp6e.h...). If your code is shit, you'll end up with 10 lb. of shit in a 5 lb. Docker container.
Time to get rid of screaming all caps BS from proteomics. When you do a tryptic digest then LC-MS, you are obviously looking at protein digest ions ("digestions" if you really need to save space...).
We have a PhD studentship available!
That sounds like the price of Thermo's latest instrument (without service contract, LC system, or Proteome Discoverer. Dry pumps are extra as are whole mass measurements. Windows 10 only. Upgrades to Windows 11 when available will require new computer hardware and a service call.)
Is a picture that is 100% FDR the best message for reliable proteomics? Ha-ha
Are you recommending replacing the 1-bar ion source for a nano rebar source? You want your concrete outputs to be strong!
All values (high intensity and low intensity) need to be within the linear dynamic range of the instrument platform to be safely adjusted by software normalizations. Overloaded samples suffer saturation, underloaded samples have things go below limits of detection. There are no shortcuts in life.
Knowing how to compare pipelines and validate results are among the hardest tasks in proteomics. If you know how to do that, you don't need any help coding. Supervisors understand time spent coding. They don't understand time spent thinking/creating. Free time AIs create will be used against you.
All aboard the noise elevator! It only goes up! How reliable are values below the limit of quant after software boosting? How far above the quant limit does data have to be for software adjustments to have validity? This is an extremely dangerous way to approach measurement science.
That is technically known as a data "poopline"...
Great. Thermo is going to transition to Windows 11 instead of Linux. I can have CoPilot hallucinate some instrument parameters, acquire scans to OneDrive where every 4th scan is an ad, and search my data with Bing. That is, if any runs finish without a security update reboot in the middle...
Now that table of contents figures are all AI-generated crap can we just get rid of them? Please. I want to scroll through the TOC to see the article titles and authors, not the shit pictures.
This SUNDAY, Sunday, sunday. It is BoxCar versus FAIMS in the monster truck showdown of the century!!! Bring the whole family. It's a night of non-stop signal-to-noise without any signal...
Can we figure out how to combine spacial proteomics with laser tattoo removal to provide population-level health care screening for millions of people? I know, the Mann Lab probably had a preprint but it might be fading over time...
ByteMe seems like a better company name...
Here is one to add: turning novel graphics into graphic novels (how AI artwork is turning scientific journals into comic books).
The heat map is the best example of why I don't do heat maps that I have seen in some time. Is that cluster-free clustering? Imagine not understanding how to analyze O-link data. O-my!
Become your own drug discovery lab!
It projects excessive false positivity...
