Densmore Shares DARPA Grant to Design New Synthetic Biology Assembly Line

in ECE Spotlight Faculty, ECE Spotlight-Research, ENG Spotlight-Research, NEWS

Assistant Professor Douglas Densmore (ECE)

Today’s synthetic biologists tend to work in silos, developing novel, biologically engineered materials and devices on dedicated platforms through sophisticated, trial-and-error experiments. As a result, new biologically manufactured products often require more than seven years to build and tens to hundreds of millions of dollars to finance. But a new, more universal synthetic biology platform is emerging that promises to dramatically accelerate the process, enabling on-demand production of new materials and devices, from biofuels to wound sealants, at a much lower cost.  

In pursuit of this vision, the Defense Advanced Research Projects Agency (DARPA) has awarded a $3.6 million, 30-month grant to Assistant Professor Douglas Densmore (ECE) and collaborators at MIT, University of California-San Francisco and Pivot Bio (a biotech startup) to help establish a “living foundry” where researchers can access, design, assemble and test synthetic genetic systems composed of hundreds of DNA parts. The new project is administered by DARPA’s Living Foundries Program, which seeks to create an engineering framework for biology that speeds production and reduces its costs by a factor of ten while radically expanding the complexity of systems that can be engineered.

The research team’s proposed method consists of three sequential tasks. The first is to create a library of more than 10,000 modular DNA parts, derived from bacteria, that would serve as biological building blocks.

The team’s second challenge is to develop an automated process to systematically assemble and use these parts to perform specific biological functions, from processing nitrogen to producing antimalarial drugs.

Densmore will be deeply involved in the third task, which is to apply this process to the production of siderophores, chemicals that bind to metal surfaces and form a protective layer to prevent corrosion, a widespread and costly (an estimated $23 billion per year) problem faced by the Department of Defense, which must operate in some of the most corrosively aggressive environments on the planet. Siderophores could be sprayed on ships, planes and other military vehicles and equipment to prolong their operational lifetimes.

“Our goal is to engineer bacteria that can create siderophore compounds in a more tuned, engineered way so that they are better performing, cheaper to manufacture and faster to produce” said Densmore, who holds a joint appointment in the BME Department. “To accomplish that goal, I will use the Eugene programming language my group has developed to create new gene clusters with machine learning techniques that use rules to bias new designs away from past failures and toward future successes.”

“Doug’s software and background in electrical engineering is critical in managing the design process and extracting actionable information from the data,” said Christopher Voigt, an associate professor of biological engineering at MIT and the project’s principal investigator.

Once the collaborators identify the gene clusters they believe will perform the best, Densmore will synthesize them using liquid-handling robots at BU.

“We envision an automated process in which people send us materials they want to design, we learn from them and improve them, and then we build new ones,” said Densmore. “If our living foundry is really streamlined, robots will test each new material and teach themselves how to build the next one.”

Informed by databases of DNA sequence information and equipped with an extensive library of DNA parts, the MIT team will build gene clusters with the potential to produce corrosion-resistant chemicals called siderophores, and test the resulting siderophores to see how well they work. Densmore will then analyze the DNA used to create the siderophores to determine rules that distinguish gene clusters that work well from those that don't, and use those rules to hypothesize how to take known gene clusters and produce better ones.