As part of our new book “What’s Your Bio Strategy?” we’ve interviewed dozens of entrepreneurs, business leaders and academics working on synthetic biology. The following is an excerpt from one of those interviews. To find out when the book goes on sale subscribe to our newsletter here.
In 2008, Jason Kelly with a group of peers from MIT co-founded Ginkgo BioWorks. The Boston-based company focuses on designing and developing organisms for use across industries and has developed an integrated technology platform that includes hardware, software, and wetware. Ginkgo BioWorks was the first biotechnology company to be admitted to Y-Combinator’s world-famous start-up incubator. An outspoken supporter of genetic engineering, Jason speaks frequently on the impact of synthetic biology. Here’s an excerpt from the conversation we had with Jason.
KARL SCHMIEDER: Is there a business that won’t be impacted by biotechnology?
JASON KELLY: Any company that has physical goods as part of its business needs to pay attention to biotech.
If you look at information technologies, every pure information business – from advertising agencies to big media companies – was impacted dramatically by digital. Every business that had a back office IT division was forced to change.
Biotech can serve a general purpose for manufacturing in the physical world the way computers are general purpose in an information-rich world. As we continue to get Moore’s law improvements across biotech, the physical part of every business will see disruptions.
JOHN CUMBERS: Did you ever think twice about starting Ginkgo Bioworks versus going to work for any one of Boston’s many biotech companies?
JASON KELLY: Today, what is called biotech really means biotechnology for developing medicines. Boston has many biotech pharmaceutical companies but when I was completing my Ph.D., I wasn’t focused on discovering drugs or studying human physiology and disease.
When we started Ginkgo, we were focused on the tools for engineering biology, improving the genetic engineering of biology. there wasn’t a fit for that at any Boston-based biotech company.
JOHN CUMBERS: What engineering principles are you applying to biology that haven’t been applied with biotech previously?
JASON KELLY: There are two categories, microorganism design and the physical work of genetic engineering.
When it comes to the physical work of biotech – genetic engineering – we’re using software and automation to do work that traditionally was done by bench scientists. This is by no means the first application of automation in biotech. The reading of DNA – DNA sequencing – for example, is a highly automated biotech process. The writing of DNA – DNA synthesis – is on its way to being highly automated.
With automation you trade off the flexibility of a process for the ease of scaling when you automate it. Automating a process that does a lot of things is harder than a process that will make the same thing every day.
The great thing about DNA sequencing and DNA synthesis is there’s a lot of exposure to widget making. At the end of the day, as long as you prepare it right, the DNA from a lot of different organisms looks the same. It goes through the same chemistry, the same processes, so you can scale that. The same thing is approximately true for writing DNA.
When you start to genetically engineer organisms, suddenly you have hundreds of protocols. If you want to insert DNA into the genome of an organism and you want to grow cells under certain conditions, you have to have scientists working at a bench. That is actually the hard part.
What’s unique about what we do at Ginkgo is, we’re applying engineering principles to standardize and automate the lab work of genetic engineering.
The breadth of the work that we’re automating is new.
We’re not eliminating bench scientists, we’re automating a lot of the repetitive tasks so they can focus on design.
KARL SCHMIEDER: One of the production bottlenecks in biotech has been measurement? Are you measuring the speed of the design, test, build cycle the way an agile tech start up would? Do you compare that to what’s being automated?
JASON KELLY: Let me talk about measuring the success of organism engineering, then I’ll tackle how we measure our work in the foundry.
Tom Knight, one of Ginkgo’s co-founders, likes to say biology is nanotech that works. It really is nanotech.
As a result, you’re operating at an extremely small scale. We’re making changes to DNA. We can read it and we can write it. It’s digital.
When you change the code, that’s reflected in the production of a gene that’s converted to a protein. For example, an enzyme that is a functioning nanomachine that is very difficult to see. That’s a general challenge in biology. You make a change to a genome, so how do you see the impact of that change?
The measurement technologies we need are extremely sophisticated. We use the latest technologies, like mass spectrometry, to get a picture of the changes we’ve made.
We’ve made huge investments into high-end analytics to give us a better view of the changes we’ve made to the organisms. Measurement is both very important and challenging because it’s at the level of the organism itself.
Designers who change an organism’s genome need to understand the impact of those changes to inform the next round of design. If you don’t understand the impact of your design, you can’t build the next iteration, and learn.
We’re investing a great deal to increase our capabilities in measurement and we’ll continue those investments for the foreseeable future. It’s a very important area of work for a biology-focused company.
With regard to the Foundry and how it operates, what we see happening at Ginkgo is a shift from what a scientist is doing at the bench today and what they’ll be doing in 2018. This is one of the reasons why I think industries outside of pharmaceuticals really need to care about biotechnology.
The scientists are not just getting better. They’re moving to a highly automated environment. We call that the Foundry and took that name from semiconductor foundries – precise, highly automated facilities. We have biotech foundries that can complete a certain lab protocol three-times cheaper in 2018 than they could in 2017.
We’ve been applying manufacturing process optimization in an automated, standardized environment that a scientist working at the bench could never do. Once every lab protocol is performed using robotics, you can track the exact quantity of every reagent, you track every liquid transfer. You track everything. Then, you can apply statistical analysis to every protocol to find out which one works, which one didn’t. Then you apply quality control and improve your success rate.
This is run of the mill process automation. It’s standard manufacturing design optimization. The same thing has been done in the manufacturing of cars, semi-conductors and everything else. We’re taking the basic ideas of manufacturing, the assembly line – automation, standardization, measurement – and applying it to biological operations. Once you have that corpus of data, you can optimize. We’re creating our own version of Moore’s law with genetic engineering because we’re measuring processes that until now have been done with no metrics.
KARL SCHMIEDER: How do you apply type of standardization and automation to organism design?
JASON KELLY: The analogy of DNA as code holds up really well. It’s a digital code. An organism essentially compiles and uses the code to execute multiple functions.
It’s pretty messy inside cells. All the functions bump into each other. You can’t isolate protocols or software libraries from each other. But at the end of the day, it’s digital. You can read it clearly, you can write it clearly, and you can program it.
We still have a lot to learn on microbial programming, on designing DNA code compared to what’s been done in the software industry. There are a lot of lessons to learn from the writing of modern software. It’s extremely complicated and there are analogies. Not everything, because biology is different. But there are lessons.
Our Foundry takes lessons from manufacturing and semi-conductor manufacturing. On the DNA design side, we’re learning from software programming and applying those lessons to biology.
JOHN CUMBERS: What does Ginkgo look like in 20 years?
JASON KELLY: I talk about the semiconductor industry a lot. Semiconductors are found in many products. They’re in your computer, your phone, your car, your TV. We have a semi-conductor industry that is in the business of designing and producing semiconductors that end up in many other products. We think the same thing will happen with microbes.
At the end of the day, the design and manufacturing of a semiconductor looks more similar than different no matter where you’re using that semiconductor. That’s why there is a semiconductor industry.
It didn’t make sense for every industry to develop its own semiconductor design and manufactured capability. It is smarter to engage with the semiconductor industry and learn how to apply what they make to your product.
We believe there will be an organism design industry. Companies will specialize in the design of organisms. Maybe those companies will end up splitting based on the tree of life. I have no idea how they’ll split.
In the near term, Ginkgo believes that if you’re going to use an engineered microbe in a process, you’d be better off dealing with the microbe design industry, rather than hiring 30 scientists and trying to create it in your own lab. That’s not going to make a lot of sense when the economics of the foundry and our learnings about designing microbes are so much better than anything you could build in-house. It will be more efficient for you to engage with us so that we design that microbe for you.
At least that’s what we see five years from now. We’re focused on centralizing design and delivering organisms on time and on budget.
Then they can deploy those organisms as part of an industry that they know better than we do. So, in the next five years, I’d like there to be a recognition that a microbe design industry exists and you’d be silly to design your own.
Predicting twenty years from now is a lot harder. We’ll be a lot better at design and will be creating much more complicated organisms and doing all sorts of fun stuff.