Autonomous, multiproperty-driven molecular discovery: From predictions to measurements and back

A closed-loop, autonomous molecular discovery platform driven by integrated machine learning tools was developed to accelerate the design of molecules with desired properties. We demonstrated two case studies on dye-like molecules, targeting absorption wavelength, lipophilicity, and photooxidative s...

Full description

Saved in:
Bibliographic Details
Published in:Science (American Association for the Advancement of Science) 2023-12, Vol.382 (6677), p.eadi1407-eadi1407
Main Authors: Koscher, Brent A, Canty, Richard B, McDonald, Matthew A, Greenman, Kevin P, McGill, Charles J, Bilodeau, Camille L, Jin, Wengong, Wu, Haoyang, Vermeire, Florence H, Jin, Brooke, Hart, Travis, Kulesza, Timothy, Li, Shih-Cheng, Jaakkola, Tommi S, Barzilay, Regina, Gómez-Bombarelli, Rafael, Green, William H, Jensen, Klavs F
Format: Article
Language:English
Subjects:
Absorption
Absorption spectra
Algorithms
Artificial Intelligence
Automation
Batch reactors
Case Studies
Chemical synthesis
Chemistry
Closed loops
Dyes
Errors
Experimentation
Exploitation
High performance liquid chromatography
Iterative methods
Learning algorithms
Light sources
Liquid chromatography
Literary Devices
Machine learning
Molecular properties
Octanol
Octanol-water partition coefficients
Optimization
Organic Chemistry
Platforms
Prediction models
Reconfiguration
Robot arms
Robot learning
Robotics
Scaffolds
Stability
Workflow
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A closed-loop, autonomous molecular discovery platform driven by integrated machine learning tools was developed to accelerate the design of molecules with desired properties. We demonstrated two case studies on dye-like molecules, targeting absorption wavelength, lipophilicity, and photooxidative stability. In the first study, the platform experimentally realized 294 unreported molecules across three automatic iterations of molecular design-make-test-analyze cycles while exploring the structure-function space of four rarely reported scaffolds. In each iteration, the property prediction models that guided exploration learned the structure-property space of diverse scaffold derivatives, which were realized with multistep syntheses and a variety of reactions. The second study exploited property models trained on the explored chemical space and previously reported molecules to discover nine top-performing molecules within a lightly explored structure-property space.
ISSN:0036-8075
1095-9203
DOI:10.1126/science.adi1407