Walid El-Sharoud @Walidws

Will we ever be able to simulate a living cell? In theory. Molecules collide, proteins fold, and all of these things can be modeled on a computer. But there are so many unknowns, and so much compute would be required, that a mechanistic model of the cell remains a distant dream. Still, many research groups are trying to build cells "from the bottom up," mostly by stringing together mathematical equations that represent different parts of the cell. By *trying* to build an accurate model of the cell, they hope to improve our own understanding of how biology works. In doing so, they also maintain legibility, meaning that humans can understand and interpret all the equations used to construct the model. But @adamlewisgreen argues that legibility is a constraint on our models. He thinks that "human legibility, our ability to understand a system...has been limiting us." To truly accelerate biomedical progress, Green thinks that we should discard assumptions about legibility in favor of "more black box" models. Neural networks can make sense of biological data in a way that hard-coded equations cannot. This is not a new argument. A 2019 essay by Bert Hubert, called "Is biology too complex to ever understand?" makes much the same point. In that essay, Hubert writes: "There is no rule that says nature cannot be more complex than our brains can handle." Instead of trying to make sense of biology via reductionist observations and mechanisms, Hubert argues that we should just gather everything into databases "that might enable computers to make sense of what we have learned." (This vision is now playing out in many research groups.) This "black box" approach might be useful (in terms of modeling cells to cure diseases, say) but would be unsatisfying in the sense that it doesn't deepen our own understanding of the universe. And isn't that a key part of science? Is science for humans -- an act that satisfies some itch -- or rather a means to an end? I think both approaches are useful. The black box models may help us cure diseases faster by making useful predictions that our brains cannot (currently) comprehend, but the mechanistic models deepen our own understanding of how cells work.

S.cerevisiae yeast does all the best things - beer, bread, champagne, chocolate, vaccines that cure cancer.... BBC: Young women now have 'close to zero' risk of cervical cancer death after HPV jab bbc.com/news/articles/…

Thrilled to share that our paper, “Deep learning reveals antimicrobial peptides within prions,” is now published in @NatureMicrobiol @NaturePortfolio. In this work, we used AI to ask an unexpected question: could proteins best known for their role in fatal neurodegenerative diseases also encode molecules that fight infection? Prions and prion-like proteins are usually viewed through the lens of misfolding, aggregation, and disease. But biology is often more layered than our labels suggest. Our study shows that these proteins can contain encrypted antimicrobial peptides, which we call prionins. This connection did not come out of nowhere. Previous studies had shown that certain amyloid-associated protein sequences, including amyloid-β and the cellular prion protein, can have antimicrobial or host-protective activity. But until now, this had not been explored systematically across prion and prion-like proteins at scale. Using our deep-learning platform APEX 1.1, we screened 19.3 million peptide fragments from 2,897 prion-related proteins. This search identified 1,179 candidate antimicrobial peptides, moving from scattered observations to a global AI-guided search across millions of possible protein fragments. We then moved from prediction to experiment. Of 75 synthesized prionins, 59 inhibited at least one bacterial pathogen, including multidrug-resistant strains. Forty-two showed activity at 16 µM or lower against at least one pathogen. Many prionins disrupted bacterial membranes, and a subset showed encouraging early selectivity, including minimal hemolysis and limited cytotoxicity. Two lead prionins also reduced Acinetobacter baumannii burden in a mouse skin-infection model, with efficacy comparable to polymyxin B in the model tested. This is still early-stage work. We do not show that prionins are naturally released during infection or that they function physiologically in host defense, and these findings do not represent a treatment ready for patients. But the study raises a provocative idea: proteins long viewed mainly as biological “villains” may also encode useful molecular functions. More broadly, it shows how AI can become a discovery engine for biology: helping us search hidden molecular spaces, connect seemingly distant fields such as neurodegeneration and innate immunity, and uncover new starting points for medicine in places we might never have thought to look. At a time when new antibiotics are urgently needed, expanding where we search matters. Deeply grateful to @mdt_torres, Fangping Wan, and everyone who made this work possible, including @Penn, @CBE_Penn, @PennBioeng, @PennEngAI, @PennEngineers, @PennMedicine, @PennChemistry, @PennMicro, @PennPsych, and @PennSAS. nature.com/articles/s4156…

@Govindtwtt I solo'd a full DNA photoshop suite in terminal using vibecoding and I dont know python. Its over 100k lines of code and it works quite well. Took me 2 months of hard work and endless testing but i could not have done it without Ai github.com/Binomica-Labs/…

SPLICECRAFT v1.0 IS LIVE!!!!!!!!!!!! Open your terminal and type in "pipx install splicecraft" if you want to try it out, then spam "splicecraft update" often as I push updates frequently. A labor of love for the community I adore. Enjoy! 💚

There are simple equations that reveal a lot about biology. For example, given the concentration of a molecule, one can quickly calculate the average spacing between those molecules in a cell. The equation involves an assumption; namely, that a cell can be sliced into a cube, with one molecule sitting in the center of each box. Each box has a side length of d, and a volume of d × d × d, or d³. Imagine zooming out of this hypothetical cell. The whole cell has a volume, V, and contains N molecules, which means it has N little boxes. So the total volume of the cell is given by the volume of one box times the number of boxes: V = N × d³ Rearranging this equation, we get: d³ = V / N Now, the term N/V is just the same thing as density, or the number of molecules per unit volume. We can therefore flip it around and call it c, or "concentration." Rearranging once more, then, we get: d³ = 1 / c, or equivalently, d = c^(-1/3) This is a beautiful little equation! Remember that c is the concentration of a molecule and d is the average spacing between those molecules. Now we can use this to figure out how tightly packed a particular molecule is within the cell. Consider glutamate, which has a concentration of about 100 millimolar, or 0.1 moles per liter. Multiply by Avogadro's number and convert (1 cubic meter = 1,000 liters), and we get c = 6.022 × 10²⁵ m⁻³. Finally, if we take the inverse cube root, we get d = 2.55 nanometers. So each molecule of glutamate sits, on average, about 2.55 nanometers from its neighbors, which is a distance smaller than the width of a typical protein. It's fun to repeat this with other things, too. Try it for water, ATP, ribosomes, and so on. If you wanted to recreate a David Goodsell painting (see below), you could also use this equation to calculate how far apart each object in the painting ought to be!

Fun interactive science app ideas | Part 3 Played around with generating 3D biological structures and made an app to explore them interactively UI Design GPT Images 2 Code Gemini 3.1 Pro More demos ↓

The creator of Claude Code teaches more about vibe-coding in 30 minutes than most tutorials do in hours. Save this - it'll change how you build forever.

Introducing Genie 3, a generative protein model that substantially advances the state-of-the-art for binder design, increasing in silico success rates by up to 20x on hard multimeric targets. It also debuts a form of inference-time scaling unobserved in other design models. 🧵1/8

This is actually shaping up to be a really aesthetically pleasing plasmid editor. Amazing what you can do without leaving the terminal. Love seeing the github say 100% Python. Looks like I'll be able to make the May 24th deadline after all!

keyboard_arrow_left Previous keyboard_arrow_right Next

What's New in Biology (May Edition) 1. The first gene therapy for deafness gets FDA approval. 2. A new oral drug for hair loss seems to actually work. 3. Big proteins fold faster than small ones. 4. Organoids can be used to make snake venom. ... and much more. With @salonium!

keyboard_arrow_left Previous keyboard_arrow_right Next

1/ Excited to share our new paper in Science: “Toward life with a 19-amino acid alphabet through generative artificial intelligence design.” @ColumbiaSysBio @ColumbiaBME @Columbia science.org/doi/10.1126/sc… 🦠🧬🛠️🖥️💥

Does this new paper mean we can stop worrying about MIRROR LIFE? > Natural eukaryotic proteases cut mirror form peptides vita-journal.com/vita/EN/10.153…

Bacteria move around using a molecular machine called the flagellar motor that rotates faster than the flywheel of a race car engine and switches directions in an instant. After 50 yrs, scientists have finally figured out how it works. “My lifelong quest is now fulfilled.” Link⤵️

We made Caliby much easier to use! Try it out: Colab notebook: github.com/ProteinDesignL… HuggingFace Spaces: huggingface.co/spaces/Protein… pip install it and use the Python API: github.com/ProteinDesignL…

@cgeorgiaw @tamarindbio Could you please share a link to this "Comprehensive Introduction to AI for protein design"?

@cgeorgiaw @tamarindbio Could you please share a link to this guide?

Makin a bunch more stability updates and bug fixes for my TUI plasmid editor. Currently workin on a pLannotate wrapper module so you can annotate plasmids imported where the author did not fully annotate all the features. I'll eventually try to pull plasmidsaurus seq results too

People think learning AI takes months. It's really just a couple of hours. And I wrote 17 free guides to start right away: Claude 101: ruben.substack.com/p/claude Claude Code: ruben.substack.com/p/claude-code Claude Skills: ruben.substack.com/p/claude-skills Nano banana 2: ruben.substack.com/p/banana-2-3bd Claude in Excel: ruben.substack.com/p/ai-couldnt-d… Best AI for Search: ruben.substack.com/p/grok-420 1M followers with AI: ruben.substack.com/p/1000000 Claude for your team: ruben.substack.com/p/claude-for-t… No prompt saves you: ruben.substack.com/p/magic AI Slides (PPT in 2026): ruben.substack.com/p/powerpoint Set up Claude Cowork: ruben.substack.com/p/claude-cowor… Claude to sound like you: ruben.substack.com/p/i-am-just-a-… Claude interactive charts: ruben.substack.com/p/claude-charts Claude as your computer: ruben.substack.com/p/claude-compu… Claude Cowork + Project: ruben.substack.com/p/claude-cowor… You're an AI workaholic: ruben.substack.com/p/ai-holic Setup AI before prompting: ruben.substack.com/p/how-to-bette… ___ 1. Save this list for later (three dots, top right). 2. Share it with a friend by ♻️ reposting this image. 3. Subscribe to my free newsletter: how-to-ai.guide.

Ruben Hassid @rubenhassid

3 months ago

x.com/i/article/2041…

41 222 2K 3.9M 6K