Machine learning speeds rare molecule discovery with info-rich mixtures
Information-Dense Synthesis for Molecular Discovery
Machine Learning
Summary
Finding molecules with very rare and useful properties is like looking for needles in a huge haystack, and current methods often work no better than random guesses. The authors propose a new way to mix many molecules together and test them all at once, rather than testing them one by one. By designing these mixtures to hold a lot of information, their method can drastically reduce the number of experiments needed to find the right molecule. Simulations show this approach finds the best molecules much faster than popular existing techniques.
What this means in practice
- •For drug development teams: Reduce costly lab tests by screening complex molecule mixtures to find drug candidates with rare desired properties more efficiently.
- •For chemical manufacturing engineers: Design synthesis processes that produce and test optimized molecule pools to speed up tailoring chemical products with specific traits.
Authors
Kasper K. Jakobsen, Eli N. Weinstein
Abstract
Machine learning can accelerate molecular discovery by designing molecules and planning experiments. However, many scientific challenges demand molecules with very rare properties, and in this sparse setting, existing algorithms offer little gain over random guessing. We propose a method to efficiently search large regions of molecular space using algorithmically controlled stochastic synthesis. Rather than design, make and test individual molecules, we design and make complex mixtures, test them as a pool, then deconvolute the molecule-activity map. We optimize synthesis to encode maximal information. Theoretically, this approach can reduce the number of experiments required to find the optimal molecule among $d$ candidates from $\mathcal{O}(d)$ to $\mathcal{O}(\log d)$ or $\mathcal{O}(1)$. In simulation, on estimated protein fitness landscapes, it finds active molecules with an order of magnitude fewer experiments than existing Bayesian optimization methods.