GAMA Miguel Angel 🐦‍⬛🔑

1/12🧵Do you want to learn how to design proteins using AI but don’t know anything about bio? I created a free 10-lesson course on YouTube. It’s now available in Spanish (original) and English (autodubbing w/Kokoro 82M). Here’s an overview of the topics covered in each lecture :)

LLMs got FineWeb, The Pile, RedPajama, Dolma. Protein ML got per-paper supplementary tables and FTP mirrors scattered across a dozen institutions. Today we're releasing AminoWeb on @huggingface : 29 cleaned, ML-ready protein datasets, ~7.5 TB total. Sequence, structure, function, MSA, variant-effect, stability, binding. UniProt, PDB, AlphaFoldDB, ESMAtlas, ProteinGym, MegaScale, Protenix, and more. Typed Parquet. Homology-aware splits. Preserved score conventions. Full provenance per record. Protein ML scaled architectures for years while the data layer stayed fragmented. We've also shared the full curation pipeline, case studies, and observations in the companion blog post. Access the data: huggingface.co/LiteFold Read the release blogpost: litefold.ai/blog/aminoweb

@TensorTwerker 🎶 Haaland, Haaland Came to us from Germany He's here to win the Champions League Ha—Ha—Ha—Haaland, hey! 🎶 youtube.com/watch?v=cixr-7…

@miangoar HAALAND motif has a certain ring to it haha

HAALAND is even found in proteins 🤪 Interestingly, other proteins with the same fold (CATH: 1.20.140.10) also contain the same "HAALAND" motif.

@daidailoh @prajdabre 😂

@prajdabre Someone activate the signal!

Babe wake up, someone reinvented the LSTM again.

#structure" target="_blank" rel="nofollow noopener">uniprot.org/uniprotkb/A0AA…

🚨 Our paper of explaining why larger pLMs do NOT always perform better on fitness/variant-effect prediction was published. See our previous thread for the key insights, and check out the paper for the full story.

@RablarR Muchas gracias Alvar! intentaré instalarlo en mi PC

@miangoar El mejor emulador de Wii es el Dolphin Sirve para juegos de Gamecube también Funcional para teléfono, PC, xbox series y steam deck Descargar las roms de cualquier lugar estaría bien Yo uso este sitio a veces romsfun.com/roms/nintendo-… romsfun.com/roms/nintendo-…

Ya que están hablando tanto de emuladores aprovecho a recordarles que pueden jugar juegos prácticamente hasta la 6ta generación de consolas usando emuladores que funcionan en el navegador No tienes que instalar nada solo buscar el juego y jugar retrogames.onl

Doing science is like ...

2/2 I know this would be computationally expensive, but the recent trend seems to be to remove triangular attention and keep only triangular multiplicative updates, which are significantly less expensive. That could make exploring higher-order interactions much more feasible 🤔

🧵1/2 One of AlphaFold2 innovations was its ability to extract more info from the same input by using triangular attention/mult. Wouldn't it be interesting to extend the rep learned by the Pairformer to n-body interactions, rather than restricting them to residue triplets?

🧬 New tutorial video drop from @EMBL_EBI! "Similarity Metrics & Confidence Scores in Protein Structure Prediction" by Dr. Yonathan Goldtzvik- Essential for anyone using #AlphaFold 🎯 🎥 youtu.be/dVFHr0ZV4-0 #StructuralBiology #Bioinformatics #ProteinStructure #pLDDT

1/9 Interesting new paper by Kasalo, Domazet-Lošo et al. on amino acid biosynthesis & animal origins. Some of the most energetically expensive amino acids became "outsourced" during early animal evolution rather than synthesized internally. 🧵 nature.com/articles/s4146…

It's funny that Margaret Dayhoff and her team invented, more than 50 years ago, the most useful protein tokenizer we still have

@hirat62363752 Yes. Check for yourself: x.com/miangoar/statu…

@miangoar N末端領域がフォールドしてないね🌟

Exclusive: Revealed what Neymar was really looking at on his phone during the match against Norway.

@Sauers_ 🤣

Qwen.........

1/ We built a protein-structure benchmark small enough to actually experiment on, but that still provides generalizable signal. Then we let agents use it to run experiments and propose new techniques NanoFold is out! github.com/ChrisHayduk/na… (scaling ablations below)

@MartinPacesa @sokrypton @LeoTZ03 @arcinstitute @BrianHie @aditimerch @bviggiano_13 @leba1127 And they were cheaper than the thousands of tokens Claude burns 🤪

@sokrypton @LeoTZ03 @arcinstitute @BrianHie @aditimerch @bviggiano_13 @leba1127 We used to call them “PhD students” 😤

Imagine programming protein, DNA, and RNA systems like you would write computer code, or even by natural language prompting of an AI agent. @brianhie and team just made this a reality with Proto: a high-level programming language for generative biology.

Jennifer Doudna won the Nobel Prize for gene editing and went on Bloomberg to say the chatbots everyone is betting on cannot innovate at all. Every promise Silicon Valley is making about AI curing disease just hit the one person qualified to check it. She has spent her whole career inside the actual frontier of curing disease. So when she talks about what AI can and cannot do in biology, she is not guessing. She is reporting from inside the lab. Her words were blunt. She is not seeing chatbots innovate. They summarize data. They write reports. They do not come up with a brand new idea nobody has ever had. Then the interviewer pushed. So you're saying AI can't innovate? Doudna did not flinch. She does not know if it can't. She just does not see it doing it right now. This lands harder when you remember who is making the opposite case. Sam Altman says AI will eliminate disease within five years. Larry Ellison says AI will cure cancer in a 48 hour window. An OpenAI executive even floated that the company should get a cut of sales on any drug discovered through ChatGPT. Doudna answered that in two words. Good luck. Even the cancer specialists Altman is selling to keep warning that cancer is not one disease but hundreds, each needing its own cure, and that compute does not skip the years of lab work. Her reason is simpler. Biology is hard. You cannot simulate your way to an understanding of the human body. The people promising cures are the ones selling the tool. The person who actually won a Nobel building them is telling you it has not happened yet. Source: Bloomberg Originals Watch the full video on their official channel.

@MartinPacesa You're right! How could I have forgotten about this incredible work. @Patrick18287926, now you can update the vertebrate dataset :)

@miangoar biorxiv.org/content/10.648…

I'm wondering whether the Steinegger lab will predict the structures of all protein-coding sequences to create a structural database of vertebrate proteins 😜