Explore a conference talk that delves into the innovative Codon Domain-Specific Language (DSL) developed by Zymergen for efficient genome design. Learn how this powerful tool enables scientists to specify large collections of DNA changes concisely, facilitating the creation of numerous test strains for improving industrial microbes. Discover the atheoretic model employed by Zymergen, which focuses on systematic searching for functional improvements rather than relying on incomplete mechanistic understanding. Gain insights into the flexible system designed by engineers, utilized by scientists, and interpreted by machinery, opening new possibilities in DNA design. The talk covers the challenges faced during Codon's development, including code trust, simplified interfaces, learning curves, and super user requirements. Understand how Codon has been successfully implemented in production for over a year, continually evolving to meet new use cases identified by scientists.
Overview
Syllabus
Intro
Talk Outline
Industrial fermentation: nature's chemical plant
The total design space for biology is enormous
The Zymergen approach
Microbial engineering: Building DNA
Microbial engineering: Building microbes
Testing microbes
Custom software requirements
Software for DNA Search: Our problem
Our solution: Codon
DNA Data Model
DNA is annotated
DnaSpecification Example
Promoter swapping
DnaSpecifications are ASTS
ASTs let us preprocess structure
How'd we do?
The "Code trust" problem
Lessons from Code trust
The Simplified interface problem
Lessons from "Simplified interface"
The "Learning curve" problem
Lessons from "The learning curve"
The Super User problem
Lessons from "super users"
Conclusions
Taught by
Strange Loop Conference
