Title: Actively Manipulating Chemical Information for Molecular Data Storage and Chemical Computing

Abstract: Living systems make simultaneous use of electrical, chemical, and mechanical domains for information processing, sensing, actuation, energy, and memory. While some of these behaviors have been mimicked in electronic systems, it is impossible for an electrical system to fully recreate the massive parallelism and emergent properties that come from the coexistence of thousands of unique chemical compounds. Inspired by these possibilities, we have been pursing methods that combine active manipulation of chemical information with more established sensor technologies and CMOS microelectronics. Molecular data systems have the potential to store information at dramatically higher density than existing electronic media. Many valuable experimental demonstrations of this idea have used DNA, but nature also uses smaller non-polymeric molecules to preserve, process, and transmit information. The theoretical limit for molecular information is two orders of magnitude denser by mass than DNA, and we have been building a suite of new theoretical and experimental tools to consider information storage using mixtures of small organic molecules. Our experimental approach combines software-directed chemical synthesis with mass spectrometry and supervised learning to write and read molecular datasets. We have encoded numerous digital files into mixtures of natural metabolites as well as synthetic small molecule libraries. Writing abstract information in chemical form also allows us to consider new possibilities for chemical-domain computations. We hope that by treating chemical systems as abstract and mutable stores of information, we can uncover new ways to interact with the natural world.