Pablo Galaz-Davison

standarized stuff is a great thing to have in science. I have never used biorender, but I find it useful if people use it wisely to generate images that convey clear messages. And it helps if everyone recognizes what is they're drawing.

The writer, drawer and reader were all limited by the same thing: their eyes. Nowadays you can generate plots that no zoom will save, or color palette. Also, the unnecessary amount of data...

well now I feel better about my dissapointingly ugly looking gels 😅

I only have a dozen articles in which I have been involved in drafting, and when I have wanted to change the title mid peer review a few times, and people have told me it is a bother and just not worth it.

I think the problem is that for us, the title of an article is meaningless, but for society its the entirety of the findings. I feel like the article title should be put at the end of the review process, as often times the article changes so much from subm to proof.

I love how this is a nice example that AF pLDDT is not strictly related to stability nor expressability.

the 15% is insane... well it's not far off the 10% we give in other places, but still. Tipping culture is always a shock. Whenever I visit a new place I have no idea when to tip and how much is a non-insulting amount.

First walk in Lyon, guided by DAMM group members at en ENS Lyon. Just amazing and what a nice bunch of people 🥰

I hadn't noticed everyone in the field that has had spaghetti predictions, in chilean Spanish "tallarín", has already settled on the term. Even though there are different kind of tallarines, some have more information than others... cool

I survived my first jet lag by pulling a one all nighter. Next day was so tired that I could sleep right on schedule 🫡

you gotta be joking THIS IS AMAZING

it was a science article? I forgot that. Fabulous. I will definitely save that for teaching

a prof (Richard Garratt) gave us once an example of a publication that literally threaded an entire structute the wrong end. So they went C to N in the density 🫣 and that was published. So yeah, human stuff happens

I think I suggested accepting donations to Sergey once (because of the beerware license), you guys would be loaded by now 😅 I know a lot of people using either your code, server, tools or notebooks, we value you <3.

This is you input the MSA and check the predictions in AF2 or use the unified framework coevolution analysis? I would like to check for our own MSA pairing strategy

that's the case for enzymes where simple words can explain it. But binders, modulators, scaffolds, motors etc, may require a bit more of a physical description of the need. In any case, new or old, the point is to design things that work and I share your view of a future with easy design.

My point is, function appears to be a rather anthropocentric parameter to categorize something that is chemical physics in nature. So it seems a bit tricky to make de novo, from scratch, desde cero, proteins with tools that replicate natural sequences then, as they may not be very novo.

In 2016 someone in a talk at the latin american protein society conference mentioned, as Kresten points out, you can have many non natural enzymatic products via unspecific catalysis of ancestral enzymes and a bit of mutagenesis. These activities depend on luck, but exist.

I think from first principle and having low identity to known protein aren't mutually exclusive features. After all, those principles are learnt from the known pool of sequences and structures. I haven't confidently used de novo, have no idea how to use it. From first principles yes, surely.

of all the possibilities, the dystopian reality we got is "unchecked capitalism enables online misinformation". So many writers turn in their grave.

this looks ideal for developing a frustratometer 👀

Well, if we dropped jargon and made sure that articles were clearer and direct instead of presumptious, maybe everyone (including ppl outside scientific community) could easily understand the main findings and individual contributions of each article.

oh this is so cute 🥹

saved it, thanks. Gonna do something similar for my current project. Still wishing for Bluesky to implement archiving posts, have dm'd myself many so far.

I see, been trying to replicate this with prompts alone and ended nowhere near. Thanks, I am going to certainly use some cat protein videos for lectures and meetings ☺️

omg this is beautiful! what was the prompt? "loving cats unfold proteins by stretching?" gonna send it to my bf

Sometimes authors summarize their responses to 3-6 major points raised by the reviewers in their resub cover letters or at the top of their rebuttal letters. This is so super helpful to editors, because it brings us right up to speed after not having read the paper/reviews for months. THANKS!!

Thanks for sharing, these are really cool, can I use them in lectures citing you as source?

yup, totally agree. I just was reinforcing the idea that codon usage can indeed have an effect on structure, just not as predictably as the original paper claimed. I accepted the original paper's result as possibly true until this came out, and just now we can safely say that it is unlikely true.

I remember there is an article in which one of such mutations altered the conformation of an ion channel, and they explained it through the codon usage. Can't find it, just this doi.org/10.1016/j.jm...