Loading video...

Video Failed to Load

Go Home

Language models understand natural proteins. But can they generalize beyond, to design completely new proteins from scratch? New preprint: A 🧵

TomSercu's profile picture

Tom Sercu

2,876 subscribers

275,213 views • 3 years ago •via X (Twitter)

EducationScience & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

🚨New paper! Generative models are often “miscalibrated”. We calibrate diffusion models, LLMs, and more to meet desired distributional properties. E.g. we finetune protein models to better match the diversity of natural proteins.
0:10

🚨New paper! Generative models are often “miscalibrated”. We calibrate diffusion models, LLMs, and more to meet desired distributional properties. E.g. we finetune protein models to better match the diversity of natural proteins.

Brian L Trippe

20,562 views • 9 months ago

We're excited to announce #AlphaFold 3 with Google DeepMind in nature: our new AI model for predicting biomolecule structures with unprecedented breadth and accuracy. Expanding beyond proteins to tackle DNA, RNA, small molecules to fuel advances in biology & drug design 🧵
0:19

We're excited to announce #AlphaFold 3 with Google DeepMind in nature: our new AI model for predicting biomolecule structures with unprecedented breadth and accuracy. Expanding beyond proteins to tackle DNA, RNA, small molecules to fuel advances in biology & drug design 🧵

Isomorphic Labs

297,005 views • 2 years ago

AI has two modes in drug discovery. Accelerate: moving faster through the existing playbook. Unlock: opening frontiers that weren't possible before. Excited to announce Profluent is partnering with Eli Lilly, the global pharma powerhouse, to unlock breakthrough medicines for patients. It's a big deal beyond the numbers ($2.25B + royalties): we’ll get to use our frontier AI models and foundational datasets to design proteins focused on large gene insertion, a therapeutic moonshot. Proteins govern almost everything in biology. We've built a generalizable AI platform to design all proteins. Onward!
0:19

AI has two modes in drug discovery. Accelerate: moving faster through the existing playbook. Unlock: opening frontiers that weren't possible before. Excited to announce Profluent is partnering with Eli Lilly, the global pharma powerhouse, to unlock breakthrough medicines for patients. It's a big deal beyond the numbers ($2.25B + royalties): we’ll get to use our frontier AI models and foundational datasets to design proteins focused on large gene insertion, a therapeutic moonshot. Proteins govern almost everything in biology. We've built a generalizable AI platform to design all proteins. Onward!

Ali Madani

43,964 views • 2 months ago

David Baker, the 2024 #NobelPrize laureate in chemistry, has achieved the seemingly impossible feat of building entirely new kinds of proteins. In recent years, one incredible protein creation after the other has emerged from Baker’s laboratory. They range from new nanomaterials where up to 120 proteins spontaneously link together (see animation) to proteins that function as a type of molecular rotor. Animation: ©Terezia Kovalova/The Royal Swedish Academy of Sciences
0:10

David Baker, the 2024 #NobelPrize laureate in chemistry, has achieved the seemingly impossible feat of building entirely new kinds of proteins. In recent years, one incredible protein creation after the other has emerged from Baker’s laboratory. They range from new nanomaterials where up to 120 proteins spontaneously link together (see animation) to proteins that function as a type of molecular rotor. Animation: ©Terezia Kovalova/The Royal Swedish Academy of Sciences

The Nobel Prize

66,375 views • 8 months ago

What would it look like to see all ~13,000 proteins inside a single human cell at once? We share ProtiCelli — a generative model that makes this possible. Besides benchmarking, we demonstrate utility in many tasks. Preprint: A thread 🧵👇
0:16

What would it look like to see all ~13,000 proteins inside a single human cell at once? We share ProtiCelli — a generative model that makes this possible. Besides benchmarking, we demonstrate utility in many tasks. Preprint: A thread 🧵👇

Emma Lundberg

38,158 views • 3 months ago

2024 #NobelPrize laureate in chemistry David Baker has succeeded with the almost impossible feat of building entirely new kinds of proteins. In recent years, one incredible protein creation after the other has emerged from Baker’s laboratory. They range from new nanomaterials where up to 120 proteins spontaneously link together… Animation: ©Terezia Kovalova/The Royal Swedish Academy of Sciences #NobelPrize
0:10

2024 #NobelPrize laureate in chemistry David Baker has succeeded with the almost impossible feat of building entirely new kinds of proteins. In recent years, one incredible protein creation after the other has emerged from Baker’s laboratory. They range from new nanomaterials where up to 120 proteins spontaneously link together… Animation: ©Terezia Kovalova/The Royal Swedish Academy of Sciences #NobelPrize

The Nobel Prize

219,336 views • 1 year ago

2024 #NobelPrize laureate in chemistry David Baker has succeeded with the almost impossible feat of building entirely new kinds of proteins. In recent years, one incredible protein creation after the other has emerged from Baker’s laboratory. They range from new nanomaterials where up to 120 proteins spontaneously link together (see animation) to proteins that function as a type of molecular rotor. Who will be awarded the Nobel Prize in Chemistry this year? We will find out on 8 October. Animation: ©Terezia Kovalova/The Royal Swedish Academy of Sciences
0:10

2024 #NobelPrize laureate in chemistry David Baker has succeeded with the almost impossible feat of building entirely new kinds of proteins. In recent years, one incredible protein creation after the other has emerged from Baker’s laboratory. They range from new nanomaterials where up to 120 proteins spontaneously link together (see animation) to proteins that function as a type of molecular rotor. Who will be awarded the Nobel Prize in Chemistry this year? We will find out on 8 October. Animation: ©Terezia Kovalova/The Royal Swedish Academy of Sciences

The Nobel Prize

65,069 views • 9 months ago

Biggest medical discoveries of the week (🧵) #1 - Evo: generative AI for biology From scratch, Evo could: Design a functioning CRISPR-Cas enzyme (top right figure) Predict gene essentiality at genome-wide scale (bottom left figure) Generate genome-length DNA sequences, encoding a diverse set of proteins (though not quite able to support life! - bottom right figure) Study:
0:19

Biggest medical discoveries of the week (🧵) #1 - Evo: generative AI for biology From scratch, Evo could: Design a functioning CRISPR-Cas enzyme (top right figure) Predict gene essentiality at genome-wide scale (bottom left figure) Generate genome-length DNA sequences, encoding a diverse set of proteins (though not quite able to support life! - bottom right figure) Study:

Samuel Hume

79,284 views • 1 year ago

A crab was seen feeding on dead skin from a person’s foot. As natural scavengers, crabs will eat almost any organic matter they come across, including dead skin and tiny parasites. Human skin cells contain proteins and nutrients, so to a crab, dead skin can simply seem like an easy food source.
0:12

A crab was seen feeding on dead skin from a person’s foot. As natural scavengers, crabs will eat almost any organic matter they come across, including dead skin and tiny parasites. Human skin cells contain proteins and nutrients, so to a crab, dead skin can simply seem like an easy food source.

Nature is Amazing ☘️

415,050 views • 1 month ago

Protein–ligand cofolding models are getting incredibly powerful… but do they have to be so slow? 🧬🐢💊 Our new preprint introduces a new flow-map framework called DeCAF for fast few-step cofolding — up to 5× faster while preserving sample quality on the SOTA Pearl model and 20x faster than Boltz 1x. ⚡🧵 📜 Blog: 🔗arXiv: Code (coming soon):
0:11

Protein–ligand cofolding models are getting incredibly powerful… but do they have to be so slow? 🧬🐢💊 Our new preprint introduces a new flow-map framework called DeCAF for fast few-step cofolding — up to 5× faster while preserving sample quality on the SOTA Pearl model and 20x faster than Boltz 1x. ⚡🧵 📜 Blog: 🔗arXiv: Code (coming soon):

Joey Bose

21,371 views • 1 month ago

couldn't resist to try creating a new UI design completely🫠
0:17

couldn't resist to try creating a new UI design completely🫠

Arnie (Milo) ✨🌛🌞

41,553 views • 6 months ago

Cells can pass proteins and RNA to each other through tubes that act as a channel between their membranes.
0:13

Cells can pass proteins and RNA to each other through tubes that act as a channel between their membranes.

eLife - the journal

192,241 views • 3 years ago

Been eating my proteins
0:16

Been eating my proteins

𝙄𝙩’𝙨 𝙏𝙞𝙛𝙛𝙮 𝙏𝙞𝙢𝙚 ꨄ︎ ₊ ⊹

16,983 views • 11 months ago

Next Ghostty Tahoe PR, adding icons to menu items as recommended by the new design guidelines. A small detail, but all the small things to help Ghostty feel right at home in the new design language when we get there.
0:13

Next Ghostty Tahoe PR, adding icons to menu items as recommended by the new design guidelines. A small detail, but all the small things to help Ghostty feel right at home in the new design language when we get there.

Mitchell Hashimoto

40,969 views • 1 year ago

🌍 Physical AI, Large Geospatial Models, and OVRMaps LLMs understand language. But the next evolution of AI is about machines that can move, see, and operate in the physical world. This is Physical AI, and it’s already happening. 🧵👇🏼
0:19

🌍 Physical AI, Large Geospatial Models, and OVRMaps LLMs understand language. But the next evolution of AI is about machines that can move, see, and operate in the physical world. This is Physical AI, and it’s already happening. 🧵👇🏼

Over the Reality 🌐

100,941 views • 11 months ago

Proteins are dynamic. These dynamics animate life.
0:11

Proteins are dynamic. These dynamics animate life.

Prof. Nikolai Slavov

45,925 views • 2 years ago

How can agents understand the world from diverse language? 🌎 Excited to introduce Dynalang, an agent that learns to understand language by 𝙢𝙖𝙠𝙞𝙣𝙜 𝙥𝙧𝙚𝙙𝙞𝙘𝙩𝙞𝙤𝙣𝙨 𝙖𝙗𝙤𝙪𝙩 𝙩𝙝𝙚 𝙛𝙪𝙩𝙪𝙧𝙚 with a multimodal world model!
0:20

How can agents understand the world from diverse language? 🌎 Excited to introduce Dynalang, an agent that learns to understand language by 𝙢𝙖𝙠𝙞𝙣𝙜 𝙥𝙧𝙚𝙙𝙞𝙘𝙩𝙞𝙤𝙣𝙨 𝙖𝙗𝙤𝙪𝙩 𝙩𝙝𝙚 𝙛𝙪𝙩𝙪𝙧𝙚 with a multimodal world model!

Jessy Lin

107,491 views • 2 years ago

Today Unstoppable Domains is unveiling a domainer's new best friend. A powerful new way to search, buy, and manage domains - using natural language and our ChatGPT plugin. You can use the tool to build & deploy custom landers with your voice:
0:16

Today Unstoppable Domains is unveiling a domainer's new best friend. A powerful new way to search, buy, and manage domains - using natural language and our ChatGPT plugin. You can use the tool to build & deploy custom landers with your voice:

Brady

23,521 views • 3 months ago

if you’re confused about how large language models work this new animation from google deepmind should clear a lot of things up
0:18

if you’re confused about how large language models work this new animation from google deepmind should clear a lot of things up

anton 🇺🇸

874,061 views • 2 years ago