Articles Featured Interviews — 09 June 2014

Jake Beal began working on synthetic biology as a side project while working on his Ph.D. for Professor Gerald Jay Sussman at MIT. He was entering the program while Ron Weiss, who is now a Professor at MIT, was finishing his PhD in the same group.  They, like others in group, were working on getting reliable engineered behavior out of lots of poorly understood, loosely behaving entities, and Weiss and others in the group made important early contributions to synthetic biology.  Beal, drawing on ideas from natural systems, worked on making abstractions that would help make better computer programs.  Eventually, however, he came full circle back to biology and started applying his work to the field of Synthetic Biology.

Dr. Beal was nice enough to take time to talk to me and I will share with you our conversation. The following are mostly his words, with approved changes for readability:

Q: Are you going to be presenting at the 6th annual International Workshop on Bio-Design Automation (IWBDA)?:

“Yes. We submitted recent work from a DARPA (Defense Advanced Research Projects Agency) project that focuses on new approaches to vaccination. Specific to that project, we are working on a recent and not well understood platform for synthetic biology, which is an RNA replicon. People have been working on it for a few years. How it works is you take an RNA virus, keeping the part that replicates and replacing the part that allows it to infect with your desired engineered payload. It is self-amplifying: a little goes into the cell but it makes a lot inside. It is an RNA system, which means that is has no direct path to affect the DNA of the cell. This makes it extremely desirable for a wide range of therapeutic applications. Essentially it injects, expresses, and then gets cleaned out.”

Q: What would it be used to treat?

“Novartis has been working on using this for vaccinations. We are thinking about ways to increase efficacy, for example by timed expression of various signals to the immune system, which might get the memory building response for the antigen that is desired, but less inflammatory response.

Other ideas are to have it go dormant and come back. That way you can have a vaccination give its own booster shot internally. A self-dosing vaccination like this will operate based on a well designed biological circuit. This sort of thing is great for the developing world where a doctor or person giving out the vaccination might only see the patient once in their life.

The trick is that if we want to do this engineering successfully, we need good predictive models. We at BBN and the team over at the Weiss lab have done a lot of lab experiments in order to get enough data to model behavior on a computer program. We have gotten to the point where I can invert the model and say this is the behavior we want and get the replicon mixture that is needed for such behavior.”

Q: Who is we?

“The folks at BBN and MIT’s Weiss lab work in harmony. BBN does analytics and the Weiss lab does wet lab, including actually carrying out the experiments. It is important to have a tight wet-dry synthesis. Dry-lab dictates a lot of more of what goes on than you might think, and wet-lab serves as a much needed sanity check and provides the scientific grounding that justifies the models that are drawn up.  Folks in the wet lab understand protocols and constraints, biological capabilities and constructs. The dry lab side is important for understanding abstraction and modularity. This is something that wet-lab folks do not have a lot of training in, just by the nature of their background. This is not a criticism, any more than it is a problem that I cannot carry out, say a PCR reaction. I work on computer science, electrical engineering, abstraction and design. So when working with the wet lab the questions are: How good of a model do you need? What do you need to know about a system to make an accurate prediction? What are the quantitative requirements that the wet lab needs for it to be useful?

This is what distinguishes synthetic biology as a field from genetic engineering. The real difference is this idea that we can raise the level of our game to describe these systems in a more abstract manner, so that we can engineer much more rapidly. We need to get out of the nitty gritty of every sequence of every system. The secret is finding out what details are not important, in order to keep things as abstract as possible. This is done by figuring out the details independently from each other and then piecing it all together.”

Q: What  goes on at BBN and what else can someone learn about or talk to you about at IWBDA?

“At BBN, we have a very broad variety of different projects, from biology to electronics, which all focus on high level abstractions. In our work with the Weiss lab, there was recently a Nature Methods paper that came out with a new family of regulatory devices that came out of the CRISPR system. That paper is about applying CRISPR to create a modular architecture for high performance repressor devices. This gives us potential to get a lot of regulatory elements that would be strong enough and similar to each other in dynamics and response to cellular context. This means it should be much easier to design reliable interconnecting control systems.

We also ran a project a couple years back that put the idea of high-level design to the test and built an end-to-end tool chain, which is similar to building computer circuits in the VLSI (very-large-scale integration) world. We were able to put the same program into both E.coli and mammalian cells, getting the equivalent behavior from each. This was like a ‘hello world’ for synthetic biology. It would fluoresce one color or the other depending on what molecule you detect.”


After discussing IWBDA we continued our conversation on synthetic biology. I have highlighted some of the points of our conversation and what I learned below.

Right now in the conventional idea of biology there is the thought that every detail is important, which stems from systems biology history. In biological systems, there are many complexity phenomena, which to Dr. Beal are often not necessary to understand. What is important to understand is that any given piece of an evolved biological system may be right on the edge of misbehavior. This possibility of chaos is not how engineered systems work, and tends to scare engineers.

The secret that Dr. Beal says we are learning about biology is that there are places where you get enough cooperation to design a reliable enough system.  Dr. Beal explained as an example that the translation dynamics from RNA to protein appears to be a lot more important than the replication dynamics for understanding constitutive expression patterns in replicons.  That means a lot of that information can be cut out and turned into one number: how much of an initial delay is there before enough replication has happened that we have seen the major effects of translation.

This is a long-winded way of saying that when you are interested into engineering the behavior of the system, your goals are driven by the engineering. For this, we need a deeper understanding of engineering design and modeling.

Dr. Beal also enlightened me on three principles that we need in order to make universal reliable engineered biological system.

  1. Get single cell data. If we don’t have that, we can’t tell if it is one cell doing a lot of work or a bunch of cells doing a little bit of work. We need information about the distribution of behaviors in the population of cells.
  2. Need to translate measurements into real units that are standardized throughout the field. This will allow for other scientists to replicate experiments. Many papers we see have arbitrary units (a.u.) which means essentially whatever came out of the analog-to-digital converter, divided by some other number to get a value that feel nice to work with. This is fine for systems biology because you can see that X is three times larger than Y, but this is not enough to engineer with.
  3. Enough experimental data to back up all of the parameters of a model.


At the end of our conversation Dr. Beal emphasized in current research investment, there is a lot of rush to applications.  Dr. Beal is all for applications because they save lives, make money, and help the environment, but we run the risk of seeing a bubble burst if we focus too much on getting applications out the door quickly. Beal stresses that we need to continue to invest in the foundations of engineering modeling, characterization, and design. This will bring visions into everyday impact.


About Author

Jacob Kurzrock

Jacob Kurzrock is a recent graduate from the University of California at Davis where he earned a bachelors of science in the field of Biological Systems Engineering. As an undergraduate he had a wide range of exposure to different labs and companies from the Plant Reproductive Biology to Amyris and then to CleanWorld. For his senior design project he helped design, build, wire, and program a strawberry harvest aid robot. In his free time you might see him brewing beer, playing basketball, or just hanging out on the beautiful beaches of Tel Aviv. On his return to California in fall 2014 he will be looking for a job so give him an offer before someone else does.

(0) Readers Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>