My toxic trait is that I believe if I looked at enough protein structures I would be really good at designing enzymes

MIT Course announcement: Machine Learning for Computational Biology #MLCB25 Fall'24 Lecture Videos: https://t.co/Puzq59oni4 Fall'24 Lecture Notes: https://t.co/RRNRY1SZaZ (a) Genomes: Statistical genomics, gene regulation, genome language models, chromatin structure, 3D genome

Anyone know which of the smaller (non-reasoning) models have good biological domain knowledge and don't cost a ton? Haven't tried out GLM-4.5 or K2... maybe they'd do well? Should I wait for new Gemini?

I made an interactive enzyme simulator. Change the temperature, add or remove inhibitors, and tune the probability that enzymes bind their substrates. Watch as the reaction rate moves up or down in real-time. Maybe this will be a helpful teaching tool.

A novel bacterial protein family that catalyses nitrous oxide reduction

Think I could make this into a screensaver or something? I catch myself watching some of these for a long time.

There's a pretty wide range of parameters, and it's all grounded in math that, for the most part, is beyond me (something about a Hopf glass and Wigner d-matrices projected into S^3 from S^4 and rotating in the higher dimension).

Been playing around a lot making generative art with various LLMs, and thought I'd start posting some. This is something I originally coded up with an experimental model on lmsys arena ('zenith', likely gpt-5? idk). I recommend fullscreen to see this properly.

Huge 3D Printing Breakthrough! Caltech is printing metal atom by atom. A hydrogel scaffold soaked in metal salts becomes the blueprint, the gel burns away, and oxygen is stripped in a hydrogen-rich chamber, leaving alloys crafted with surgical precision. They can change metal

Codex with gpt-5 is actually goated, I'm kind of astounded. I don't feel like I have to watch it like a hawk like with Claude to make sure it doesn't delete half my code or hard-code tests to pass. I love Claude but wow it's like a weight lifted off my shoulders

If this work is generalizable, it would be transformative. Computationally guided genome rewiring of Escherichia coli and its application for nanopolyethylene terephthalate (PET) biodegradation and upcycling @TrendsinBiotech

🧬🤖 ESM2‑guided GAN engineered 12-mer peptides whose ΔG-ads is ~30 % stronger than PepBD’s best. 🎯One binder for two plastics (PE + PP) at once, a first for microplastic cleanup. Which pollutant should AI tackle next?

https://t.co/WVsxRzqMeZ

🧬→🧩 Sequences gave us AlphaFold. 👥→🌐 Proteomes now give us ProteomeLM: attention heads self‑discover PPIs, trimming interactome mining from 30 days on 100 GPUs to 10 min on one card. I think this is the first transformer that reasons over an organism’s full proteome 👀

https://t.co/lgkQAif6nL

🌀Turns out, fold‑switching seems common🌀 ~4‑5 % of globular proteins in E. coli🦠 It's also evolutionarily selected, and largely invisible to vanilla AlphaFold

We are thrilled to announce our new publication in Science Advances: Gaia, an AI-powered protein search platform that brings genomic context into functional annotation. https://t.co/nNNogNBk07 Gaia enables rapid, scalable discovery of remote homologs across 131,000+ genomes —

The ChimeraX 1.10 release candidate and daily builds can predict small complexes of proteins, nucleic acids and small molecules using Boltz on your Mac, Windows or Linux computer. https://t.co/ffSIINo04Y

https://t.co/kqb4V788SN Recognize evil when it shows its face to you.

Sometimes I think that proteins are truly motherfuckers like this one that binds Uranium ☢️💀 Protein with twin binding sites for uranium extraction from seawater https://t.co/lOOsVhikrB img source https://t.co/SCYSh1pO3W