Why Joyn Bio Won’t Build a High-Throughput Fermentation Lab
Joyn Bio is engineering microbes for sustainable agriculture. Joyn’s Head of Bioprocessing Kelly Smith, PhD and former CEO of Pasteuria Bioscience, talked to us about what it takes to build a biotech startup, and why that does not include an in-house bioreactor lab. In the interview, find out why she won't build a high-throughput fermentation lab at Joyn.
Culture: Before we talk about your role at Joyn, it would be helpful to understand your background and the experience you bring to what you do now. I noticed your PhD is in environmental engineering. What made you interested in this topic and what was the focus of your research?
Kelly: I've been interested in environmental protection most of my life and always wanted to have a positive influence on the world. The department I studied in at Caltech was heavily involved in water resource protection and air pollution mitigation. One of my professors designed the wastewater outflows for the nuclear power plant at San Onofre, so that's the kind of thing I thought I was going to study. Then I got interested in microbiology because one of the professors in the department was a microbiologist, and she was studying the use of microbes for remediating hazardous waste.
I did a postdoc with the Smithsonian looking at nitrogen cycling with microbes in soils as a nitrogenous pollution mediation strategy. That led me to work in the paper industry for a company called Hercules Chemical in their division that helps paper mills control microbial growth. It seems like a stretch, right? To start in microbial ecology by remediation, and then to go work in the paper industry? But paper mills are full of water, heat, carbonaceous inflows of wood pulp, and recycled cardboard, so they actually have interesting and unique microbial ecosystems. We studied what kinds of microbes made up those paper mill ecosystems, and then what the mill could do to control the overgrowth of the bacteria without discharging a lot of toxic chemicals into the environment or harming the quality of their products.
From there, I went to work for a startup called Entomos because I wanted to get involved in all aspects of developing a product. As a researcher in a big company like Hercules, I didn’t have many options: I could stay in the lab and do routine work for a long time or go into management. Instead, I decided to work for a startup so that I could be involved in the early research, product development, launch, and fundraising. That was how I got interested in agriculture. I barely knew a seed from a plow at the time, but in going to work for that company I learned about the pest control industry and about how microbes can help meet the need for better pest control solutions in agriculture.
Culture: What was your experience like working for a startup?
Kelly: I actually quickly went from employee at one startup to CEO of a different one. Entomos was in the middle of raising their Series B round of funding when 9/11 happened, so the capital markets shut down for months. Eventually they decided to shut down most of the business except for the division I was working on, which was using microbes to control nematodes. That division then became its own company, Pasteuria Bioscience. That’s how I became a startup founder. A little different than the typical path, the CEO of Entomos came to me one day and said, okay, you're going to be the CEO of the new company that we're starting. And I said, okay, I guess I'll go raise some more money then.
Culture: What was it like going from working in a lab to running a company?
Kelly: I was fortunate to have good mentors. The location of the new, three-person startup was a 550 square foot lab in an incubator run by the University of Florida. It was specifically started to commercialize biotech inventions and to give little biotech companies a place to start. This incubator also had programs for helping scientists become business people, so I got a lot of training. Experienced entrepreneurs and retired business people listened to my pitch for fundraising and gave me feedback. They coached me on how to talk about tranches and return on investment and things like that, which I understood mathematically as a scientist and engineer, but you have to learn how to speak that language. I also had great mentorship from an organization called Springboard Enterprises, which focuses on helping women become entrepreneurs, so it was a great intersection of support I had at the time.
Culture: So you grew the company from three people? What was the trajectory from there?
Kelly: The three-person startup is a great story now. We had one small lab and I didn't have an office - I just had a computer on a bench in the lab. When I had calls with our Board of Directors, I had to go stand in the hall. It was all we could afford. It also happened to be where they stacked up the bins of biohazard waste to get taken away. It was the worst lab in the building, but it was the only one available when we needed it. In the end we had between 35 and 40 employees, and the company was sold to Syngenta. I worked for Syngenta for about a year and a half after the acquisition, and then I went to AgBiome who had a new venture with the soil microbiome. I saw a big opportunity there to expand the number of biological products that are available to solve growers' problems.
Culture: Given your experience with all those different types of products and companies, what are some of the unique challenges of biomanufacturing and fermentation you’ve seen within the agriculture space?
Kelly: There are two major challenges with biomanufacturing for ag. The first one is that it's still a relatively new idea. Large ag companies have a strong history of producing chemicals for pest control. There are still questions about whether a biological product will work, whether growers will buy it, or whether customers accept the fact that we're putting bacteria on their crops.
The other challenge is the cost. Ag products have to be inexpensively manufactured for them to be competitive, especially against some of the older chemical products that are generic and cheap at this point. A lot of the contract manufacturing that’s available to pharmaceuticals is just too expensive. Some of that cost is driven by the fact that for pharma you need GMP-quality facilities and people who are trained in that. Ag products don't have that level of regulation, but there aren't as many CMOs that are equipped to work with an agricultural product.
Culture: How does the need to drive down cost affect your approach to bioprocess development?
Kelly: We have to look for microbes that grow fast. When you're doing any kind of fermentation-based manufacturing, the biggest cost driver is how long the main fermentation takes. That's where your labor cost comes in, your energy cost, and to some extent your raw materials cost. We try to optimize the fermentation process so that the fermentation can run as quickly as possible. The other big factor is how much it needs to be processed at the end of the fermentation. After you harvest it, every additional step (e.g. handling, formulating) adds costs to the final product, so you want to minimize the number of downstream processing steps that your product needs.
Besides looking for fast-growing microbes, you should also be looking to reduce the cost of your media. Even yeast extract gets to be a bit of a cost problem at scale. You want to work with microbes that can be adapted to grow in a cheap type of broth. Look for inexpensive sources of sugar or choose a manufacturing site where an inexpensive source of sugar is already available. Also, select for microbes that already have the properties you want. It’s a lot more expensive to try to adapt the microbe and make it do the things you want in the fermenter than to select for one that has most of those properties naturally.
Culture: Speaking of adapting microbes, are there still concerns in agriculture around making genetically modified organisms?
Kelly: Sometimes. Depending on your market, it's easier to register microbes that are not genetically engineered, and you'll have less resistance to adoption with a non-GM product. At least in the U.S., there's a rapidly growing understanding that the biggest impediment to microbial products getting adopted is more about efficacy and cost, and people are getting used to the idea of genetic modification as a way to make the product do what you want it to do - as opposed to something that's scary and unnatural in some way. I think that message is getting through now - that genetic engineering is a useful tool, but there are definitely markets where people will not accept genetically modified products.
Culture: If you’re working with organisms that are not as “standard” (like E. coli or P. pastoris), is there a challenge in trying to culture something that’s never been cultured before in a bioreactor?
Kelly: It's difficult to predict how any new microbe host will behave in a fermenter. Then, in going from a glass to a steel fermenter and then to larger steel fermenters, you encounter many variables that are difficult to model or predict. There's still some intuition and art involved in scaling up a novel microbe, but there are also some generally applicable strategies to try.
You can do things like start with a standard panel of carbon sources, for example, and do some statistical design of experiments to look at different levels and combinations of carbon and nitrogen. Generally, if it doesn't grow in a little shake flask culture, it's going to be hard to scale it up to fermentation, and it'll take a while. So, when you’re looking at lots of different microbes from the environment, you want to pick the ones that already grow a little bit in screening studies.
Culture: What about challenges around data management and analysis? Especially with DoEs where you’re dealing with large datasets, what are some challenges with analyzing data and using it to make decisions for scale-up? And have those changed over time throughout your career?
Kelly: Right now, especially with technology like what Culture Biosciences has developed, we have the ability to test every hypothesis, so the question on the front end is, which hypotheses are really worth testing? The question is no longer which ones can we test? With these capabilities and with the automation and robotics that are available now, the experimental space that you can explore is much bigger. Managing the data that comes out of that has a lot of very familiar challenges. You've suddenly got a huge volume of data, but how do you turn the data into information? You need experts in the design of experiments and the statistical models that you use to help make sure that the results you're generating are statistically significant and that they’re based in physical reality, not just an artifact of the testing. All of those “big data” kinds of challenges come to bear because the field of biomanufacturing now has the ability to generate a lot more data than we ever could before.
Culture: What led you to Joyn - where you are now - and what are your responsibilities there?
Kelly: I started at Joyn a little over a year ago to build the bioprocessing and formulation team. Up to that point, the fermentations were done by our parent company, Ginkgo Bioworks. When it became apparent that Joyn needed its own fermentation team, they recruited me to come in and build the team. I handled the first contract production of material for the field trials last year and built the seven-person team we have now. I also started the regulatory group because there was a growing need for things like permitting.
The great thing about Joyn is that we can now make the microbe into the product that we want. Finding a naturally occurring microbe that has all the properties you want for your agricultural application - and can be scaled up, manufactured, formulated, packaged and sold cheaply enough for agriculture - is pretty rare. Instead, using synthetic biology, you start with a microbe that has some of the characteristics that you're interested in and then make it do the other things. I found it to be an unbelievable opportunity to get these products into the hands of growers who desperately need them - and to do that more quickly and efficiently because we can genetically modify the microbes now.
Culture: Having built teams in the past, what was your vision for how to build your team at Joyn?
Kelly: As a manager, I focus on how to keep people engaged with their work, and I hire people who are intrinsically motivated to do it. The intrinsic motivation is what keeps people working through the tough problems, and there are a lot of those in research. There's a cartoon that I love, which shows this ream of unpublishable results next to one little paper that finally got published - anyone at a research-oriented enterprise can relate to that. People have to be intrinsically motivated to keep with the problem and keep testing the right hypotheses until they get the results that we need to get a product out.
At my last job, I learned about research that’s been done on what makes people stay motivated and engaged at work. It comes down to three things: autonomy, mastery, and purpose. People need to be as self-directed as they're capable of being; micro-managing people and looking over their shoulders is the quickest way to cause burnout. Mastery is the idea that people feel good when they learn new skills, and it makes them want to do better work when they feel like they're learning something and getting good at it. Providing people with opportunities to develop mastery of a new set of skills, whether it’s lab skills or interpersonal skills, is critical. The third thing is purpose. Agriculture’s purpose is fairly straightforward - feeding the world, and keeping a supply of safe, affordable food. Making sure people can see the connection between what they do day to day and the larger impact of that work is important to keeping people engaged.
Culture: How have those guiding principles influenced how you set up your lab infrastructure as it relates to bioprocessing and bioreactor capacity?
Kelly: I don't hire people whose only job is going to be something tedious and routine, at least not for very long, because that quickly leads to burnout. Anything that can be automated should be automated. Running big banks of small bioreactors can get to be tedious and labor intensive, so it’s a better use of Joyn’s resources to let a specialist like Culture Biosciences handle that. We have some internal capacity, but only for specialized purposes.
For the general routine screening and what we refer to internally as a design-build-test cycle, it makes sense to have Culture Biosciences execute that because that’s Culture's main focus.
Culture: When you first reached out to us, had you considered whether to build a bioreactor lab in-house and hire a team to run it?
Kelly: Honestly, that wasn't ever a question for me. Knowing that the bioreactor infrastructure already existed at Culture in your cloud lab, I thought we should just use that. As an engineer, I try to avoid putting steel in the ground, as they say. Because Culture is doing this kind of work as its business model, I felt like you guys are the experts at running bioreactors and we should partner with you and take advantage of that expertise.
Culture: What types of projects have you worked on with Culture so far?
Kelly: Almost all the work that we’ve done with Culture has been on our flagship product, which is a microbe that will fix nitrogen. It will become a replacement for a lot of chemical fertilizer. Our target is to replace 30% of the fertilizer applied to corn with these microbes.
Culture: What kind of impact has Culture had on your work, especially compared to what you would have seen if you’d built out a bioreactor lab internally?
Kelly: We’re several months ahead of where we would be if we had tried to build this in-house because it would have taken a long time to source the equipment and get it installed - not to mention finding an affordable place to put it. Lab space is at a premium in Boston, as you can imagine. Finding the lab space, sourcing the equipment, hiring the people to run it, getting it up and running, and validating the work would have taken six months to a year. I think we're at least that much time ahead of where we would have been if we had tried to build a lab ourselves.
Culture: What would you say to a potential customer of Culture’s that is considering building a lab internally? Are there advantages to building an in-house lab?
Kelly: The biggest advantage to building your own internal capacity is having complete control and 24/7 access to it. I think that's a decision that a young company has to weigh - whether the considerable capital investment, the time it takes, and having to hire and manage the personnel to run that facility, is worth it. I generally don't think it is worth it. With a company like Culture, you get total control and transparency anyway with the ability to monitor experiments in real time. The benefits of not having to manage all that infrastructure are much greater than the benefits of having your own facility.
Culture: How has working with Culture shaped the way you do your R&D planning? Where does Culture fit into your overall process?
Kelly: Right now we’re in what we call the design-build-test cycle. The engineered microbes get tested in vitro and the ones that pass the in vitro screen come to us. Those are the ones that we send to Culture to grow in large enough quantities that we can look at the growth characteristics and the fermentation profiles that you provide to us. We can see how they compare to each other and to the wild type strain that was used as the host. That material is then sent to our facility for on-plant testing in greenhouses. We do this as an iterative process, where we take the output from all of that and use it to inform the next round of engineering and the next batch of strains that comes from the engineering scientists.
Working with Culture has been great for this iterative process because it’s so easy for us to get the data that we need to inform the next experiment. Culture’s web interface for data visualization and analysis is clear and really easy to use. Everything we need is there. The data is accessible in whatever format we need to have it, whenever we need it. Our data science team has been in touch with Culture’s team directly to understand how we can even more efficiently use that in our models.
Culture: Are there certain features of Culture’s cloud-based software that are the most useful to your work?
Kelly: The willingness to work with us on the API - that's huge. The fact that the data, as it's generated in the fermentation lab, is already in a form that we can use? That's gold. With most of the other labs I've worked with in the past, the data is generated by off-the-shelf bioreactor infrastructure in a proprietary format. Most of the output that the operators see is just time charts and they don't know how to get it to us in a format that allows us to use it as input for our models or store it in an internal database. They’re usually not set up for that. As a result, we’d have to write software internally that would translate the data into the database we're trying to build. On the other hand, working with Culture, we don’t need to do that because the data is already there, ready for us to use.
Culture: How do you see your relationship with Culture evolving over time?
Kelly: I think it’s the beginning of a beautiful friendship. The kind of routine strain screening and iteration that we need to do seems to fit well with what you're offering as a core business. I'm looking forward to doing a lot of work with Culture.
Culture: Finally, what advice would you give to early stage biotech companies that are just getting started?
Kelly: I think it's important for a young startup to focus on building a good team with the right skillset. Conflict management is an underrated skill that's really important to prioritize when you’re building your team. Anytime you have two people in a room, inevitably, there's going to be conflict, so your choices are to manage it well or manage it poorly. The more people come in with that skill, or the willingness to learn it, the better off your team's going to be.
On the technical side, you want people with a solid grounding in microbiology and the math that goes into understanding bioprocessing. I think a lot of chemical engineering programs in particular focus on that right now. And when I say solid grounding in microbiology, I mean not just knowing how to make clones in E. coli, but actually growing a bunch of different kinds of microbes on different types of media, knowing what they are supposed to look like when you streak them out on an agar plate and how to tell whether a culture is contaminated. That kind of experience is vital to a bioprocessing-based startup.