Max Fürst

Comp. Bio. preAI:
spend days finding (=missing many), understanding (=skimming paper), installing (=missing libs -> bang head), running (=curse hpc queue), & interpreting(=🧐🤔🫣😵‍💫) output of scientific software

postAI:
spend days trying to talk an AI agent into doing the darn thing for you

#1 lesson I give to protein designers for wet lab purposes : Please design with a Trp and/or multiple Tyr. Please do not design with 10 cysteines.

Remember when we were on social media because heck how else would we find cool stuff like this

I set myself a reminder to check this 1 year later & turns out I was mostly wrong🥲
Here search results from on "DNA", last year (orange) vs now (blue) and the indexing issue is only noticeable for the current year ('24), barely an increase for '23
New hypothesis: all (bio?) topics peaked around '21?

arXiv will no longer accept review articles and position papers unless they have been accepted at a journal or a conference and complete successful peer review.

This is due to being overwhelmed by a hundreds of AI generated papers a month.

Yet another open submission process killed by LLMs.

wouldn't it make much more sense to plot this relative to normal distribution found across the proteome?
Too lazy to extract the numbers and make a scatter plot, but I think Ala and Glu are at least as off the axis as the aromatics
www.uniprot.org/uniprotkb/st...

Re recent AFDB update, in case you wondered:
- most of AFDB is still same original predictions
-new/changed entries were modeled with AF2
- the MSAs are the originals, so should not contain sequences from last few years

Hi Maxim, thanks. I see that only changes were newly predicted and these were indeed made with AF2 again, not AF3.
However, not clear to me yet if the provided MSAs (for unchanged entries) are the originals from 2021ish or if you recomputed

Pretty vague, given the bombast. Since the only numbers I see there aren't apples to apples, can you contextualise for us whats now the actual change in $/genome (or some other quantitative metric)?

Not sure the backwards reasoning here is apt? IMO BindCraft does stand out *because* of large number of tests (12 protein targets by my count). They may still have been unusually lucky / biased in target selection, but maybe it isn't the most obvious example to illustrate ur very valid overall point

Upper half of a PhD position accountment in the field of biochemistry.

Lower half of a PhD position accountment in the field of biochemistry.

I am looking for a PhD student interested in enzymes!

For the full annoucement: www.chemzymes.com

I wrote a blog post about the future of structural bioinformatics.

Where to go after AlphaFold? How do we avoid the field becoming a load of half-baked LLMs?

Let me know what you think.

jgreener64.github.io/posts/struct...

YouTube video by decay21450 You Don't Own Me - Bette Midler, Goldie Hawn & Diane Keaton

youtu.be/Q_oFL_b719g?...

also, this editorial's answer to the question is yes (in violation of Betteridge's law), recommending to just don't produce AI slob that doesn't add anything

can't find the article now, but I saw a very useful graph once w/ # of articles vs # of reviews in a field:
-many articles, few reviews: write review
-few articles, many reviews: write review of reviews
-many articles, many reviews: write specialized review
-few articles, few reviews: write proposal

It's becoming increasingly clear to me that Reflect Orbital's fucking stupid giant mirror satellite, with absolutely NOTHING useful to offer, which will cause countless safety issues, ecological disasters, and destroy the night sky, is going to launch.

A bunch of astronomers and I have sent out […]

Groen licht voor grote AI-fabriek in Groningen. Europa investeert 71 miljoen euro De kogel is door de oude tabaksfabriek: Groningen krijgt een grote AI-fabriek. De Europese Commissie gaat 71 miljoen euro investeren. Het benodigde bedrag van 200 miljoen is daardoor bij elkaar.

Groen licht voor grote AI-fabriek in Groningen. Europa investeert 71 miljoen euro

Woohoo! Go Foldit!

Maxim could you kindly clarify: did you actually recompute all the structures, or are the models identical to the previous AFDB version? And are the MSAs the ones from back in day, or do these reflect more recent db searches? If new predictions were made - with which AF version?

Great post! In agree that complexity is underestimated, lack of knowledge the bottleneck to design, & intuition key to success. Likely, AI can work work around some, but not all of the issues without explicit data.
Do you mind sharing what pockets you think work better than others with bindcraft?

Have you found the entire db MSA data as a dump yet?

Ooohh providing the MSAs is wonderful. Will the data be in a new Google cloud bucket? Is it already available?

We’re renewing our collaboration with Google DeepMind!

We'll keep developing the AlphaFold Database to support protein science worldwide 🎉

To mark the moment we’ve synchronised the database with UniProtKB release 2025_03.

www.ebi.ac.uk/about/news/t...

🖥️🧬 #AlphaFold
@pdbeurope.bsky.social

For sure. I'm not downplaying the breakthrough itself. But I do think the other, dependent achievements would not never have come. How much later - who knows.

Figure 1

Figure 2

Figure 3

Mapping the evolution of computationally designed protein binders [new]
Evolve comp. mini-protein binders *in vivo*. Maps seq-affinity.

Agree they should get the Nobel, but if it hadn't been that synthetic route, another one would've eventually come, good enough at least to make primers. Or we would've dugged earlier deeper into primases or protein priming. Oh wait, guess maybe we wouldn't have been able to study those without PCR 🤔

Un-frigging-believable. I have waited for something like this for so long!

Wait but that's inconveniently more complex and (😱) mathy, my 1st years will not be happy

@science.org ⤵️challenges prevailing notion—repressor and activator act by distinct mechanisms: repressors block via steric hindrance/activators enable Pol II binding to promoter🧬

E. coli transcription factors regulate promoter activity by a universal, homeostatic mechanism | Science bit.ly/adv2064

The AlphaFold Database Ages

Figure 1

Table 1

The AlphaFold Database Ages [updated]
AlphaFold DB:
...needs updating as entries conflict with newer UniProt releases, showing bioinformatics databases' rapid obsolescence.