Platform
We experiment beyond limits to design smarter. Our integrated experimental platform is built for generative AI-driven drug discovery and lets us tackle unsolved problems.
Exploring Beyond Limits
The number of small molecules that can be made is so large it’s functionally infinite. So how do we search it?01
Diversity Library Screening
02
Focused Library Screening
03
Generative Drug Design
Overview
The right blend of experimentation and computation, powered by the highest quality data.
Our experimental platform is built for generative AI-driven drug discovery, integrating chemical experimentation and computation on an unprecedented scale. We explore molecules and targets more broadly and deeply with a sophisticated combination of ultra-high throughput experimentation, generative AI, biology, medicinal chemistry, automation, and nanotechnology.
01
Design
Generative AI leverages a growing database of 2B+ target-ligand binding measurements02
Make
Adding millions of new library molecules a week to a collection of 60M+03
Test
Screen 2M molecules against a target in four minutes04
Analyze
Convert 25 TB of image data into binding affinities daily, allowing us to rapidly identify potent and selective moleculesWorkflow
An iterative workflow designed to leverage the full power of modern computation.
How it works
Experimentation
tArray
Make encoded libraries on nanobeads
Molecules are combinatorially synthesized on nanoparticle beads and DNA barcodes are attached to identify which molecule is on each bead.
Immobilize beads on tArray chip
Beads are randomly placed into 32M wells on a chip the size of a nickel.
Map position of beads
DNA barcodes are sequenced to map the position of each molecule, then DNA is cleaved off the beads.
Measure target-ligand binding at scale
Precise binding affinity is measured based on fluorescence intensity in less than 4 minutes per chip.
Resynthesize and test additional properties
Hits are resynthesized at microscale and further evaluated using high-throughput biological testing.Computation
tCompute
Data processing at scale
25TBs of image data are processed and converted to actionable target-ligand binding data every day.
Understanding the language of molecules
Multi-modal contrastive pretraining for representing and traversing chemical space.