ray__

I can obviously see its value if there is no window manager or desktop environment on the system I'm using, but that's rare these days–if I'm working on a system with neither of these it's always over ssh. With a WM or DE it seems like tmux just adds a redundant layer of organization. What am I missing?

BioRender is cool for premade vector images, but their license (and cost) is still too prohibitive for that, especially for grad students. Bioicons (https://bioicons.com) is a nice FOSS alternative.

“I’ll give you a lesson, a half hour for $20. I have some children that come just to see me once a week and I give them a lesson—$20 for a half hour. And there’s a lot of NYU students that come by, we give them a discount for being students. One hour for 40 bucks.”

This post could have easily been written about alcohol in the prohibition era. We've since learned that criminalizing alcohol makes its impact on society worse rather than better. We can't strip drug users of their autonomy and their ability to lead any sort of normal life and then act surprised when they turn to crime or turn back to drugs when they have nothing left.

With Zotero, syncing reliably across multiple devices was always hit or miss for me. Saving a few papers throughout my day, going home and reading/annotating them on my iPad, and then going back to work and pulling up a paper from the previous day with its annotations intact is the dream workflow for consuming academic lit that has always just worked with Papers but always required a lot of effort with Zotero. Maybe this is more of a "me" problem than a Zotero problem, but if I can't configure Zotero to sync reliably then many academics won't be able to.

There are some other neat features too. The built-in reader is good enough to not bother going with a third party. The handling of in-text references is better than any alternative (the actual reference is listed when hovering or clicking the reference number in text–a seemingly basic feature that lots of citation managers and PDF readers don't have that is essential for maintaining concentration). The handling of figures, metrics, and smart suggestion of related articles is also top-notch. I was sad to make the conversion from FOSS but haven't looked back yet. Hopefully this new Zotero release will make it possible to transition back–it looks like lots of killer features in Papers are offered by Zotero 6, which is great to see.

As for the rights-holding thing, most journals give their authors the right to distribute their work for free upon request, and there are some other loopholes for getting papers afaik, so I don't think it would be worth the time or effort to go after people with a few sci-hubbed papers in their Papers account. I am afraid that DRM will make this much harder though, especially for papers that have been downloaded on a university network for later offline reading.

If it were purely genetic, evolution would have selected against these genes long ago since T1D was a terminal disease with a life expectancy of less than a year until the discovery and development of insulin-based therapies in the 1920s.

Its algorithm seems to work really well for rapid consumption of buzz within a particular network that you can tap into by following people. It is one of the few cases where I like the fact that I occasionally get out-of-network Tweets in my timeline because it usually exposes me to something really interesting posted by someone outside of my direct field. I'd never want to use Twitter for comprehensive reviews of some topic, reliable back-and-forth discussion, etc. but it's just fantastic for gossip and trends.

The tables here are even starting to turn–so many people are on Science Twitter that it feels like I find as many new and relevant papers on Twitter as I do in my tailored email notifications and RSS feeds. Twitter is also always a day or two ahead, and the papers usually come with the backstory and context in plain English. I hate to say that the Twitter thread from the paper author is often more informative (at least, per unit of time that it takes to consume) than the content of the paper itself.

Solid tumor penetration isn't really related to this though, it has a lot more to do with the fact that it is physically difficult for a molecule to diffuse through the many layers of cells that make up a solid tumor. When you take a drug, it generally ends up in your bloodstream and from there must diffuse through the lipid bilayers that encapsulate cells (whether they be cancer cells or not) in order to reach their target. This diffusion is a big barrier when it comes to designing drugs, because most things won't passively diffuse through lipid bilayers. A successful small-molecule drug will be able to 1) bind to its target effectively enough to stop that target from doing some disease-causing thing, 2) not bind to other things that are important for cellular function, and 3) get into the cell in the first place, without being broken down before it gets there. Balancing all 3 of these requirements is tricky, but rules of thumb have been developed for 3) that help guide the design of small molecules.

Perhaps the most important guideline for 3) is size. Most small molecule drugs (anything that you take in a pill, along with many chemotherapeutics) are designed to be < 500 Dalton. Once you get over 800-1000 Da diffusive cell penetration is rare (there are interesting outliers, cyclosporine cruises through lipid bilayers despite weighing in at ~1200 Da). Immunotherapy generally involves retraining your immune system by introducing antibodies (~150 kDa+) or whole T-cells. These modalities can generally only target things on the outside of cells, because there is no way they're getting inside, and they certainly won't be able to pass through the many layers of cells that make up a solid tumor.

tl;dr is that immunotherapeutic agents won't be able to penetrate solid tumors by diffusion because they (the antibodies and cells involved in immunotherapy) are too big, and there isn't any other mode of entry. I do wonder if a true immune response would need to penetrate at all though, because presumably T-cells would break down a solid tumor layer by layer if the appropriate antigen was present. I'm not sure how correct this line of thinking is though.