95% of rare diseases have no FDA-approved treatment. The traditional drug discovery model is broken for 300M+ patients. 🧬
Eratos can fix that:
🔗 https://t.co/xVT6zjTObm
#RareDisease#AI#Eratos#DrugDiscovery
1/4
Exciting news from the team! 🧵
Our latest preprint is now out on bioRxiv!
Using our AI-driven ASCEND™ engine, we analyzed spatial data from Alzheimer’s models to uncover previously unreported nanoscale protein associations.
3/4
By combining AI with high-dimensional spatial biology, we’re able to uncover molecular relationships that may otherwise be difficult to identify—and potentially reveal new druggable interfaces.
The full paper is now available on bioRxiv.
If you're interested in World Models, spatial biology, BioAI, drug discovery, or computational therapeutics, take a look.
👇 Full paper:
https://t.co/t5Hzba1WXU
#WorldModels#SpatialBiology#JEPA#YannLeCun#DrugDiscovery#CNS
1/10
What if a model of biology could learn not just what molecules are, but how they organize in living tissue?
That’s the idea behind HI-JEPA, Eratos’s World Model of molecular organization.
And its training data are different from most biological foundation models. 🧵
9/10
This is what we mean by a true World Model of Biology:
Not simply describing biology.
But learning biological organization well enough to:
- simulate it
- plan over it
- identify emergent biology
- design therapeutics against it