Using AI to predict biochemistry for manufacturing and therapeutics
The next decade is going to see a biotech revolution fueled by three technologies: ability to read DNA (sequencing), ability to write DNA (synthesis), and computational systems that predict what DNA to read and write. With sequencing and synthesis being mainstream now [1], 20n provides the computational systems that predict DNA design for novel industrial biotech and health applications. At 20n, we are taking a fresh look at turning biological data into information. We approach it as a big data learning problem, and bring to bear techniques from machine learning, distributed systems, AI planning, program synthesis, data mining, and natural language processing. With the support of DARPA [2], we are building a team of computer scientists, and bioengineers to look at biology in fundamentally new ways. Our current customers include industrial biotech firms that are leading the charge in transitioning us from traditional unsustainable means of manufacturing to sustainable biological processes leading to advanced fuels, materials, therapeutics, and flavors and fragrances. 20n’s technology helps them enter markets previously thought inaccessible for biology. [1] Sequencing DNA costs $4k/genome (http://www.genome.gov/sequencingcosts/) and is expected to go down to $1k/genome. Synthesizing DNA is at 20cents/bp, and is expected to go down to 0.1cents/bp. [2] DARPA in the past supported work that led to the internet, stealth planes, GPS, and even Douglas Engelbart’s mother-of-all-demos.
