we help you launch discovery programs with a query.
La Jolla
Joined January 2024
- Tweets428
- Following16
- Followers14
- Likes3
1/ 🧪 named pharma + pathology AI: @AstraZeneca × Tremont AI (announced Oct 27; covered Nov 5) will develop and evaluate AI for preclinical toxicologic pathology safety screening. research collab only. 👇
5/ 📊 related: VivoSim × Chemveda (Nov 3) is another AI→human-relevant tox loop, on 3D liver/intestine NAMs instead of digitized rat slides. ir.vivosim.ai/news-releases/…
6/ ✅ primary: tremont.ai/news/tremont-ai-a…
TRACE paper: biorxiv.org/content/10.1101/…
if you read tox path: would an encoder like TRACE change triage first, or only second-read on rare findings?
1/ 💊 named pharma AI deal: @Roche will use @chaidiscovery's AI models and platform in a number of drug discovery programs across therapeutic areas (announced Nov 5). 👇
5/ 📊 related: named AI×pharma deals keep stacking in 2026. BMS × Boltz was an earlier stack-in-the-loop pattern. full Chai announce: chaidiscovery.com/news/roche…
6/ ✅ primary: chaidiscovery.com/news/roche…
if you run a biologics discovery portfolio: would you bring Chai-3 onto a challenging target first, or keep it as a parallel design pass?
1/ 🧪 AI designs → human tissue: VivoSim Labs × Chemveda (Nov 3) will put AI-designed compounds into VivoSim's 3D human liver and intestine models. Chemveda designs/makes; VivoSim tests. 👇
5/ 📊 related: FDA has been pushing human-relevant NAMs over animal tox. and @AstraZeneca × Tremont AI (Oct 27) is another named pharma bet on AI for preclinical safety readouts. tremont.ai/news/tremont-ai-a…
6/ ✅ release: ir.vivosim.ai/news-releases/…
if you screen AI-designed small molecules: would human 3D liver/intestine models change which compounds you advance, or only confirm animal tox?
5/ 📊 related: RNA tertiary prediction has been the hard gap next to protein AF3. and yesterday's @chaidiscovery × @Roche deal is another named AI×pharma push into designable biologics. chaidiscovery.com/news/roche…
6/ ✅ paper + code + data:
nature.com/articles/s41586-0…
github.com/heqin-zhu/structR…
if you work on RNA structure or ncRNA function: would you start from structRFM weights, or keep a sequence-only LM?
