Genesis Final Round: Nanyang Biologics
summary
Proprietary model DTIGN predicts 3D compound–protein interactions and claims 27% more accurate bioactivity predictions versus leading methods; pipeline fine tunes open source LLMs to extract from >20 million papers and processes 1 million papers in about six days. Built a natural library of 10 million compounds, runs a 6,000 sq ft lab, reports $3.5M in deals this year and $6M next year, holds IP on the model, and validates discoveries including a plant compound hitting the DNA damage response that depends on farming practices.
AI-written summary · about
related talks
10mGenesis Final Round - Drylab AI
36mFormatting Biological Big Data to Enable (Personalized) Systems Pharmacology
16mBeyond Patterns: The Case for Physics-Grounded AI in Drug Discovery
29mProtein folding and what it means for drug discovery
25mBiology's ChatGPT Moment
30mLeveraging Molecular Data for Drug Discovery
30mMachine Learning: A New Approach to Drug Discovery
36mAlgorithmic Alchemy: Transforming Drug Discovery with Open AI