Strange Loop

Next: September 12-14 2019


Stifel Theatre


St. Louis, MO

Designing 1000 genomes in a single bound: The Codon DSL

Today, humans use DNA to cure diseases, pinpoint genetic disorders and create new materials. To do this effectively, scientists must be able to design and concisely specify collections of changes to DNA for in-lab testing. At Zymergen, we create large numbers of test strains when searching for improvements to industrial microbes, which means our scientists need powerful tools to allow them to search the genome easily. Zymergen's wet lab is fed input specified in the Codon language, an in-house DSL developed for specifying large collections of genome designs with a concise amount of code.

Prior synthetic biology toolkits focused on specification of individual designs, reinforcing a model of mechanistic understanding of the genome that is frequently incomplete. At Zymergen, we apply an atheoretic model, searching for functional improvements rigorously and systematically. This search algorithm involves testing similar mutations at many points in an organism's genome. The Codon DSL allows scientists to easily build up combinatorial sets of genome edits.

Codon has been in production for over a year. Our scientists continue to identify new use cases, and have even written python programs that generate Codon scripts of high complexity.

This talk will present how the Zymergen team is designing a flexible system built by engineers, used and understood by scientists, and interpreted by machinery, opening new doors to designing DNA.

Aaron Kimball

Zymergen, Inc.

Aaron Kimball is the CTO of Zymergen. Aaron's career has involved work at several big data-oriented startups. At Zymergen, he is responsible for building the data and automation architecture to facilitate a robotics-driven approach to DNA design and microbe engineering. Aaron was the first employee at Cloudera, where he helped with a variety of big data challenges. In 2010 Aaron cofounded WibiData, the big data application company. Aaron has been working with the open-source project Apache Hadoop since 2007 and is the original author of Apache Sqoop and MRUnit. He holds a bachelor's degree from Cornell University and a master's degree from the University of Washington, both in computer science.

Theron Rabe

Zymergen, Inc.

Theron Rabe is a Senior Software Engineer at Zymergen, leading development on the domain specific language Codon. Prior to this, he lead the Search Team at Epocrates, and earlier, worked on a variety of research concerning formal systems, natural language processing, and programming language theory. Theron earned a BS in Computer Science from Winona State University.