Enabling Tools for Non-Model Organisms with Henry Lee

On this episode of The Climate Biotech Podcast, we are joined by Henry Lee, CEO of Cultivarium, a Focused Research Organization (FRO) that accelerates the engineering of non-model organisms. 

Henry came to biotechnology through electrical engineering, after a chance opportunity to dissect crayfish neurons inspired a passion for using engineering to understand biology while using biology to engineer. During his postdoc in George Church's lab, he pursued what he calls "bucket list science." For one of his projects, he developed new methods to culture and engineer the bacterium Vibrio natriegens, which is one of the fastest-growing organisms ever described and grows roughly twice as quickly as E. coli. That work shaped Cultivarium's approach to the broader problem of non-model organisms. 

Henry believes that the mental model of applied biology has traditionally been far too one-to-one, focusing on a single organism for a single application: “If you’re interested in plastic degradation, well have you heard of this one microbe that does plastic degradation?” In reality, most useful capabilities are shared amongst many organisms, and we would ideally draw on that full diversity to develop solutions. Cultivarium is developing tools to make it easier to work with diverse organisms to find what works best for a particular goal. 

Listen to learn how Cultivarium is changing the way researchers approach engineering non-model organisms.

Transcript

Reusable Lessons From New Microbes

SPEAKER_00 0:03

Of all the different microbes that we're working on, what are the lessons that we're learning that are reusable? What are the principles that we can use so that when we look for the next N plus one, N plus two, N plus three microbes or even organisms in general, how do we tackle these problems? So we're trying to learn about learning.

Show Mission And Guest Introduction

SPEAKER_01 0:24

Welcome to the Climate Biotech Podcast, where we explore the most important problems in climate and environmental biotechnology and how we can solve them. I'm Paul Reginado, co-founder of Homeworld Collective. Together, we have agency to build technologies that enable a brighter future for all life on Earth. Alright, I am here with Dr. Henry Lee, who is CEO and co-founder of Cultivarium, a nonprofit bioengineering firm accelerating the pace of engineering environmental bacteria, archaea, and fungi, so their unique physiologies can be used for food, fuels, materials, and medicine. Backed by the Eric and Wendy Schmidt Foundation and Welcome Trust, Cultivarium has already reduced the time and cost of engineering new to lab microbes by more than tenfold and is building toolkits for more than 300 species, from halophiles that thrive in brine to filamentous fungi poised to reinvent biomanufacturing. Henry, thank you so much for joining.

SPEAKER_00 1:22

Thanks for the invitation, Paul. It's great to see you again.

SPEAKER_01 1:27

You have been present through a lot of my journey in biotechnology. When I first started my PhD in George Church's lab, you were right there doing crazy stuff, domesticating v naturgens, and here we are almost ten years later.

From Electrical Engineering To Ecology

SPEAKER_01 1:42

So let's just start with your background. Like, how did you get here? Did you always know that you would be domesticating strange microbes? Or was this as surprising to you as it was to everybody else?

SPEAKER_00 1:57

Both, actually. It's both extremely childish of me, and something that I never expected I would do. So, what do I mean by that? So, my background, I grew up with a lot of computers, building computers and playing around with circuits. And my undergrad degree was in electrical engineering, and I was really into building all sorts of things and taking things apart. And during that time, I had two formative experiences. I, after pinging around university as undergrad, ranging from anywhere from astronomy to physics, I thought, hey, there's this really weird thing called neuroscience. Let me go check out what that is. And that was an amazing experiment, experience where I got to dissect a crayfish, put a needle inside this giant neuron in the tail, and record electrical signals from a giant neuron. And that really ignited my love of this idea of biology and engineering being able to intercept and that we could use engineering to try to understand biological things. During that time, there was this huge renaissance or this huge boon of synthetic biology, where lots of people were talking about how cells are like computers because they have these genes and these are the parts that then can be wired. And so I thought, let me go check that out. So that's sort of a linear progression of me getting from engineering to biology. But the idea of studying non-model organisms or just organisms around us really is a childhood dream of mine because when I immigrated to the United States from Taiwan, I we moved to Arizona. And Arizona has this amazing program or amazing thing called the biosphere, where they have this giant dome, glass dome, where they're trying to put a biosphere synthetically inside a sealed place that was for you know better understanding our world, thinking about how we might establish a biosphere off-planet. And, you know, that was the time when I started thinking about wow, how would you actually build an ecosystem from scratch? What does it mean to what are the ingredients of that ecosystem? And if someone were to find an organism and say, here, you figure it out. Is this important or is this not important? Should it go in this ecosystem or not? What does that mean? So that's actually something that I think about all the time, which is what are the ingredients parts of our ecosystem? How do we better understand it? How do we understand how much of a role it plays? And so that's what drives a lot of my curiosity and my and my passion for the work that we do at Cultivary.

SPEAKER_01 4:39

I love that. That's so beautiful. I didn't know that that sort of ecological motivation was like at the core of or at the root of this work. And so that sort of that began that that was there early on for you, it sounds like, and was with you through your career arc.

Collins And Church Lab Mindsets

SPEAKER_01 4:58

And you went through two of the most prolific synthetic biology labs in the world during your training as you were on this journey, starting with Jim Collins at Boston University and then George Church's lab at Harvard. What did that experience teach you about how to build things in biology and about your, you know, how to approach this question you had about how to choose which microbe is right for building an ecology?

SPEAKER_00 5:25

Yeah, so this was quite an interesting experience, which is to say Jim Collins, of course, a giant in synthetic biology, systems biology, did a lot of that influential seminal work in this space, but he's a physicist by training, a nonlinear dynamics guy. And so when I joined the lab, I was really, really enamored with all of these how do you treat a cell like a computer? What are the parts? How do you put it together? And so it was actually a kind of a crazy wake-up call, or at least being thrown in the deep end, where I would go to these lab meetings and there would be all these biologic biology jargon that I just didn't understand what they were talking about. So, for example, and this is not something that I'm that I shy away from telling people, probably for the first six months of my grad school career, I had no idea what the word an operon meant. I didn't know you know what transcription factors really meant. So I was sitting around just really trying to absorb all of this new biological jargon that I hadn't been exposed to in my undergraduate training. So that actually allowed me to try to be more like Jim. And one of the cool things about Jim that I really admire is that he has this ability to just cut through a lot of the noise and straight to what is at the core of a problem. And so I think part of me being thrown in the deep end and learning from and observing from the way he and other members of that lab worked was to be able to say, yes, there's a lot of detail, there's a lot of stuff to know, but what is the exact question that we're trying to solve and how do we do that? And that's to some extent a you could call it an engineering approach to things, you could call it a first principles way to tackle problems. So I think that was a really great way, though painful at the time, because I thought I would come in and have this, you know, really great grounding in classic biology. But I think that not having that actually helped us think differently, dream differently. I'm not even going to say bigger, because that that that of course has a lot more nuances, but it just gave us a different perspective on things. So I think Jim's lab was really in a really formative time on trying to level myself up from someone that's coming from engineering and to be able to look at problems from a more engineering perspective in biology. So that's one piece on Jim. Now, George, of course, has a different perspective on things in that his perspective and his lens is effectively the entirety of the universe and all possible ideas. And so, you know, joining his lab, you're sort of blown away at the range and the and the courage to be able to think and at least speak about ideas. And I think one of the things that you really take away from a place like George's lab is the value of being able to talk about ideas even if they're not fully formed. The value of, I guess some would call this creative destruction, where there might be a lot of ways in which people approach problems in a canonical way or in a presumed way, and that you could then try to think differently or try to tackle the problem differently. And so that was really fun for me because I could bring some of these ideas and some of these approaches from Jim's lab into the entire catalog of George and George's lab members, and of course, with people like yourself who are we're all trying to this basically we're all trying to explore this massive universe of ideas and scientific possibilities nearly simultaneously. And so that was really great. And you know, I think maybe another character characteristic of George is that he basically will just be up for most things, at least as a prototype or at least as a trial. And so when I joined the lab, I thought this is probably the only place on earth where I can propose to work on an esoteric microbe that has nearly nothing else other than it's been described, it's been put in a strain bank, and I would like to learn from scratch, again, itching that scratch of what is this organism? How useful is it? What are its parts? How do you make it work? How does it work? And so I remember I was sitting with him in a meeting and we were talking about this, and I think it was probably a five to ten minute meeting where we talked about, okay, let's do a new organism, what might be fun.

Why Vibrio Natriegens Matters

SPEAKER_00 10:01

How about the fastest growing one that anybody has ever described, Vibrial Natrogen? So we thought, okay, let's try it. And out I went and we started working on that. Now, of course, the one last thing I would say is, you know, it's it was an extremely humbling experience working on something new, because there's just not a lot of precedent for establishing a new organism or domesticating that organism in the lab. And so I did a lot of looking through old literature, trying to come up with mental frameworks or patterns, but a lot of times it's really just plain old luck and it works. And so a lot of what we're trying to do at Culturum is to try to figure out of all the different microbes that we're working on, what are the lessons that we're learning that are reusable? What are the principles that we can use so that when we look for the next N plus one, N plus two, N plus three microbes or even organisms in general, how do we tackle these problems? So it's really we're trying to learn about learning.

SPEAKER_01 11:03

Beautiful. So maybe we could dig into like what are some of these, what have you learned about learning? And specifically, well, what have you learned about learning how to cultivate organisms? And maybe there's things you've learned about learning as well in general that you could share. But it's I think it's it's very cool that you started with one novel organism that and I'm actually curious, did you come in like wanting to work on a weird microbe because you wanted to work on a weird microbe? Or did you come in wanting to solve an important problem? And then you were like, oh, v nitrogens grows quickly, and so so we should do v nitrogens. Let's start with that.

SPEAKER_00 11:43

Yeah, I could tell you the genesis of that. So joining George's lab, I told myself this would be the one place in on earth and in my life where I would do what I call bucket list science. And so that is I'm just gonna go in, try some stuff. If it goes well, awesome. If it doesn't go well, I had the chance to just play. And I'll tell you a few things that all came to a head in how we chose Vibrio Natrogens. Number one, I've watched too many sci-fi future out outer space movies. So, you know, let's just pick on Independence Day for one. If ever there was an alien organism, I want them to call me to try to figure it out. So I thought, this is the good place. I'm gonna just pick an organism and I'm gonna trial it out. So and we'll make it simple. It'll be a it'll be a bacteria, it will go super fast because then I'll get that iteration speed, right? I'll be able to test lots of things. So that's one thing, right? So that's one piece. The second piece is the ability to grow really fast, or some of these fundamental questions on what governs that speed, what are the biological principles, or are they really are they thermodynamic? I mean, almost everything is thermodynamic at the end, but what are the limits of being able to grow fast? And that that was another thing that really captured my imagination, which is could we try to use genetics to figure out why an organism like vibranitrogens grows so fast? And by the way, I couldn't figure it out. I tried really hard in my own special way. Others are still trying hard. I don't think anyone has a smoking gun on why or how vibrant grow so fast. And so this is still an open-ended question.

SPEAKER_01 13:26

And how fast does it grow?

SPEAKER_00 13:28

Well, that is still up for a little bit of debate because the classic paper by Egon clocks it at 9.8 minutes per generation. And so that is pretty fast. But in my hands, I've never gotten it to grow much faster than about 14 minutes per generation. Whereas E. coli within the same types of conditions, 96 wall plates, you know, 37, all that sort of stuff, is about twice that at about 28, 29 minutes per generation. So the sort of framework of vibrant natures grows twice as fast as V. coli. That I've confirmed with my own hands. Other scientists have done that with their own hands, but the exact magnitude of whether it's 9.8 minutes per generation or not, that's that hasn't been reproduced as far as I understand. But by the way, let me also tell you that there is a faster organism. It's just that it's flesh-eating and it doesn't produce much of its own nutrients. It relies on a host. And so there are some really interesting and tantalizing ways in which you could work on that biology if really fast growth and the principles and limits of fast growth is your jam.

SPEAKER_01 14:35

And I had asked you this question about learning, and we'll get to that, but I actually I want to talk a little bit more about v nitrogens just to make sure listeners like understand why it's so helpful to work with an organism that has such a fast growth rate, fast doubling time, right? Because you know, growth is exponential, right? So every doubling time you get twice as much of your organism when it's in exponential growth phase. And so what does that mean for your ability, the ability of a science to do experimentation on a microbe when you have something that grows twice as fast? Like E. coli is what most that you know, it's like the first organism that you would use if you're trying to engineer some DNA, unless you have a important reason to use a different organism. And the and part of that is because E. coli grows so fast, 30 minutes is fast, but you have something growing twice as fast. And so maybe tell us about that and what it enables.

SPEAKER_00 15:37

Yeah, so at least at a first approximation, growing faster should yield much more productivity in biological research. At least when what whatever interactions, what whatever you're doing within the microbial space. So let me break it down a little bit. So a lot of times when you're trying to, of course, if you're studying bacteria, if you're using bacteria to study basic biology, the faster something grows, the more observations you can make, or the more iterations of a particular experiment you can do. So that's useful in its own right. What we use a lot of E. coli for, if you're more interested in studying other types of biology, like human biology, mammalian biology at large, is you might use E. coli to create DNA constructs. And so you're really using E. coli effectively as a little chamber that holds a piece of DNA that copies a piece of DNA, and then it just is a self-replicating machine that will make many copies of your DNA that you can then blow up and then take only the DNA that you want. So it's a DNA manufacturing or replication engine. Now, one of the one of the thoughts, the knock-on effects I thought of working on Vibrinatrogens to make it, you know, easier to work with in the lab, compatible with all the tools, is that we thought, well, if it grows twice as fast, surely as a first approximation, that means that everything gets done in half the time. And so you could do twice as many experiments, or you could go home half the day, right, in half the day. So, of course, in reality, that's not really what happens. There are certain aspects of science where you're just waiting a long time or you're going home, and it's okay to have to have something incubate overnight. So you come in and you you pick up your plate with colonies later. Now, we've actually been, I wouldn't say criticized, but I think people are still a little skeptical on whether or not adopting something like vibro natrogens will be helpful to their workflow. And by and large, if all things considered, yeah, they might not be helpful because vibrionatrogens is still not E. coli. It's not, uh it doesn't, we don't have a built-up as a group, an expectation of how it works. And so anytime something goes wrong, people, well, maybe I shouldn't have used that thing. Maybe I should have done the tried and true thing, right? So that's one thing. Number two is yeah, science is still done extremely manually by arms and by humans standing there moving things around. And so you have to do science with biology as the rate limiter, both in terms of the organism you're studying and the organism studying that organism. So you have to have lunch, you have to go home, you have to rest, but that might all change. And so I'm really actually quite excited about this new interest in autonomous labs and the ability to have automation or robotics come in and drive a lot of the more manual steps of science, because now we're talking about having science move at the pace that it can move, where a human no longer has to babysit it or chaperone it. So that's a little bit of what I think about when it comes to speeding up science. Now, there's a lot of other details that we can go into, which I've made this point before, others have made this point, including Nico McCarty, where many of the actual processes that we have for, let's say, cloning DNA, if you actually sit down and map out the amount of time that we spend with that end-to-end workflow of putting a piece of DNA together, putting it in a cell, transforming these cells, picking a plating it, waiting for it to grow, picking a single colony, letting that grow. If you actually quantified all the time that it takes to do each step, you'll actually find that a large majority, I think 70 to almost 70, 80% of a workflow can be just waiting for a cell to grow. Right. So if you're looking at it from that perspective, then yeah, an organism like Bible Natrogen, which is growing twice as fast, will save you time. And that that's what I was really keen on bringing to the scientific ecosystem.

SPEAKER_01 19:51

Very cool. And so as you were doing that work, you were learning how to learn about an organism, right? And now you are at cultivarium, you're scaling learning about organisms to many organisms, and to make it as it sounds like to me the goal is to make it more and more of a simple method than an art of exploration. And so, yeah, maybe you can tell us what you have learned about learning organisms.

Learning How To Domesticate Microbes

SPEAKER_00 20:26

Well, the first thing I learned about learn about learning is that there are a lot of learned people out there in history with their awesome papers. And so the first thing is go out and study all the papers. Now, if you could go back to the JSTORs and all these other classic papers, what you'll often find is that yes, there is some theory, but most of the reports are feminological. Wow, we did a thing and it worked. And then maybe there's a hypothesis that really you can't test if unless you're doing lots and lots of different microbes at scale, right? In order to see if your hypothesis is generalizable or if it's just specific to this one microbe. So one of the things that I live through, and that many others who have tried to work on these non-model organisms have lived through, is that oftentimes you feel like you're solving for a very specific problem for that particular microbe. Everything feels bespoke because we have tried to put our scientific process has been I care about this microbe, I must make this microbe work as if it's the only example on earth of something that is useful for this goal, right? So I think that it's important for us as we think about learning is that in order to figure out if what we're learning is generalizable, we have to test on many things. And in order to test it on many things, we also have to acknowledge that our current mapping, intellectual mapping of biology is far too one on one. And what I mean by that is oftentimes when you talk to people about biology and applications of biology, you might say something like this. You might say, you know what? I am really interested in plastic degradation. And therefore, if you're interested in plastic degradation, have you heard about this one microbe, just this one microbe that does plastic degradation? Well, we worked on that at Cultivarium. Idionella, now renamed, and you'll have to excuse me for not remembering the Latin name, but here's my point about it, which is we far too often have this one-to-one mapping of application equals a Latin name. But really what we're what we probably benefit from thinking about is an application is probably an entire set of Latin names. And maybe more importantly, it's a set of genetic pathways or metabolic pathways that can be present in a whole cohort of microbes that are relevant for what you're trying to solve, right? So this is really about stripping away a lot of the canonical thinking around how biology is classified so that we can take a more engineering-based solutions forward approach to trying to solve things. So that's one piece on that. Now, maybe what you're what you're also referring to that listeners might be interested in is well, what have we learned? Are there any key things that we've learned about what makes something really easy to grow or something really easy to transform and what are the molecular parts? And so people are doing this all over the world all the time in terms of trying to grow something. Lots of people have models for our nutrition or metabolic networks. We have built our own tool called Genome Spot.

GenomeSpot And Growth Predictions

SPEAKER_00 23:49

What it tries to do is it can take a fraction of the genome, look at how proteins are encoded by their amino acids, and it can infer or guide the optimal salinity, pH, oxygen, or temperature that an organism will grow in. So that's pretty cool. Others are taking a much more empirical approach, just trying a bunch of stuff. Now, on the other end, you have a bunch of people working on molecular tools, which means things like I'm gonna do saturating immunogenesis on every possible, every possible base of this gene, I'm gonna figure out exactly what is the optimal encoding of this gene and how it works in this microbe. So cultivarum actually has now spent a lot of its time thinking about what's in the middle, which is once you have a bunch of the cell mass, and if your goal to do modern genetics and genomics is to then get recombinant DNA in, your key problem is how do you do delivery? How do you do nucleic acid delivery? And so that's what we've started spending a lot of our energy on is how do you do that? Now, punchline, we don't know exactly how to do that in a really predictive way yet. So we're trying to collect this data and we're trying to collect it in the following ways. We have built a robotic uh work cell that instantiates as many of the transformation or DNA delivery modalities as has been published, uh, so that the idea is you can submit to that work cell a plate of cells, ideally of different species, and it will try all of them. And where there is some signal, it will actually run an iterative loop to improve it, so that you can not only discover how to do DNA delivery, but you can improve it. So that's a very practical thing that we're trying to do. The other thing we're trying to do, which is learning about learning, is where we have taken all sequenced genomes of bacteria that we can find publicly available. We've run as many annotations of traits as possible of those genomes. So you basically have per species or per strain a bunch of labels on all these different traits. And then on the other end, we have has it been transformed by this method, that method, and we're trying to create some mapping between the two. And as of right now, we don't really care if it's black box or white box. We don't know exactly what the explainability is of those relationships, but we're trying to figure out are there rules or are there predicted ways in which we can take an organism and say, now we know how to transform it. So there's a lot of these types of projects going on inside Cultivarium where we're really trying to crack this transformability problem. So I'll stop there, see if there's interest in going into different areas, but hopefully that gives you a little vignette on a bunch of different ways that we're learning about learning.

SPEAKER_01 26:54

The Climate Biotech Podcast is powered by Homeworld Collective, a 501c3 nonprofit unlocking biotech solutions for planetary health by fostering community, building knowledge, and directly supporting early stage research. We are always looking to connect with scientists, innovators, and funders who want to accelerate progress in this space, whether in our current program areas of critical minerals and greenhouse gas removal or in other application areas. If that's you, reach out to us at hello at homeworld.bio. I mean, this so just to summarize, you have you pointed out two key aspects of working with a microbe. One is how to grow that microbe, right? And the other one is how to get DNA into the microbe. Or in other words, DNA delivery. Because for any kind of bioengineering, we need to actually get DNA in, and that's one of the biggest barriers in all sorts of different kinds of biological experimentation. The part about DNA delivery sounded fairly simple. You have the list of all the ways of delivering DNA, you try them out, and you see which ones work, right? And you don't care if it's a black box. I mean, it sounds like there's some very complicated mechanisms under there, but we're not quite sure what those are yet. But the genome spot one seems to have some like it's fascinating in its simplicity, right? You were saying that you know genome spot is this, I guess, analysis method that lets you go from the amino acid frequencies in an organism to a prediction about its growth conditions. And I actually just wanted to clarify is it the amino acid frequencies or is it the codon usage?

SPEAKER_00 28:38

It's on the amino acid level.

SPEAKER_01 28:40

Okay, so this is like so amino acid frequencies is like how much, you know, there's 20 different amino acids roughly, how much of each amino acid is present in this organism. And so it might not immediately be intuitive why such a simple quantifiable metric could tell you so much about what an organism needs to live, which would seem like you know, a very complex, you know, that if it what temperature it's at, what salinity, that has all sorts of effects on physiology beyond just proteins. So do you have a sense for why the codon usage tells you so much? Not codon usage, amino acid frequencies.

SPEAKER_00 29:29

Yeah, yeah. So I really love this tool from Genospot. I think it's better to call it a tool rather than a finding because in order to make it a finding, we'd have to actually go in and perturb all of these or genetically change all of these signals to make sure that we are or catching something causal and not correlative. So from as a scientist to other scientists, we believe this is at least correlative and that we have some work to do if we want to call it causal. But here's an intuitive way to think about it. So what GenomeSpot does is it looks at the protein amino acid frequencies of different charge residues or bulky hydrophobic, so on and so forth. And if you start to think about it within how proteins are folded, this might make some intuitive sense, right? So if you have thermophilic proteins, they might have more charge residues and bulky hydrophobic so that they can form more salt bridges and tighter hydrophobic cores so that this fold is really robust when you're trying to grow at ADC, something like that, right? And maybe vice versa when you're thinking about psychophiles that really grow at cooler temperatures. If you're starting to think about halophiles and salinity, those are different amino acids and pH have different surface charges, so you you can intuitively get the sense that there are the way that you encode your proteins, which do all of the actual functions, right, the actual doing of the cell, that makes a lot of difference because you need them to be resistant, you need them to the conditions that the organism is growing in. Now, I think this is really has some interesting ramifications if you start to think about what this might imply. Right. So again, we're at the correlative stage, not necessarily positive, but you can start to you almost start to buy it when you start to think about it in this way.

SPEAKER_01 31:28

So this is brainstorming, this is hypothesizing now.

SPEAKER_00 31:30

Exactly, exactly. So, so I'll give you, for example, with vibronatrogens. I thought one of the things I spend some time doing is I thought, you know what, vibrantrogens it grows so fast, maybe there's a magic gene A or a magic gene B. And if I just shoved it into E. coli, maybe E. coli will just grow faster. Well, that was never the prevailing hypothesis, but I thought, you know what? It would really be silly if I didn't try that and there was a magic gene A or a magic gene B. And by the way, I also wanted to learn how to do shotgun cloning. So I thought, win-win, right? Try it out. Now, what's interesting is now you have people interested that are looking at how do you take a microbe and how do you make it an extremophile? How do you just engineer it? Can you just shove more things in it in terms of genes? Or can you just make a few tweaks here and there? And so what GenomeSpot was making me think about really intensely was well, if you actually wanted to create these really sophisticated or at least really complex phenotypes, like being a halophile or any sort of extremophile, that probably is a genome-wide or systematic change that needs to happen, right? And so that tells you a little bit about the lift that it might require if you're trying to do synthetic genomes or trying to turn an E. coli into something that lives in extreme conditions, or trying to take a micro from Earth and then having it work on Mars, right? So it's really fascinating to think about how wild those changes have to be and how exhaustive they have to be on the whole genome scale in order to get these remarkable phenotypes. Which then brings

DNA Delivery And Transformability

SPEAKER_00 33:17

me back to, well, gee, we don't know how to do that really at scale. So we better take the different microbes that we have on Earth and try to figure out how to work with them instead of taking the N. coli, for example, and trying to push that towards something completely unevolved for the particular condition that you're interested in.

SPEAKER_01 33:36

I love the simplicity of it. And I love the way that the simplicity of a metric like amino acid frequency then relates to something really sophisticated and complicated. So let's let's move on to another tool that you recently released called Prism. That's an acronym, PR-I-S-M. Maybe you can explain what the acronym stands for. But this is a tool kind of differs from your previous ones. It's less about discovering new ways to work with organisms, more about spreading existing knowledge, right? This is like an AI platform for recording video uh protocols and sharing them, right? Which is something people have tried to tackle in the past. I guess the new advances in AI make it more tractable. But what led you to prioritize that in light of Cultivarium's mission and how are you seeing it being used?

SPEAKER_00 34:28

Yeah, so Prism really was born out of this really simple nuisance, to be frank. So so we're interested in studying all sorts of different microbes, and we oftentimes will run into things like we'd look up papers on how this microbe was grown, and it would, you know, you look up the methods and you're pouring over the methods and you're thinking, wow, why is this written in prose? I really just need an like a set of instructions or like a cooking cooking instructions almost. And so for certain organisms, we ended up running into things like, oh, you know what? If you want to work with this organism, really what you need is you need to go visit that lab in the Netherlands of this, you know, 80-year-old guy who and that's the only place. If you're legit about this organism, you have to go train there. And uh wow, that's really romantic. And that's really you have to go to the guru. Yeah, and I kind of wanted to do that, but then you know, I logged on to Twitch and I was like, cool, that's pretty cool. Someone's like playing a game and I can watch them do a thing. And I thought, wow, what we really need is just like a Twitch for scientists. We need to watch how people are doing things, and I need to be able to transmit that in this era of YouTube, right, and TikTok. What could be simpler? And so we tried to think about okay, well, how can we incentivize people to share these tips and tricks? Because oftentimes, and you'll go to all these different labs with people working on these basically esoteric organisms, and there's all these things that you do by muscle memory that you no longer remember to write down. Because if you know, you know, and if you don't know, well, you better go train in you know the outskirts of the Netherlands.

SPEAKER_01 36:08

It's also a lot to write down. Like that's like if a picture is worth a thousand words, then like the process is like how many words is that?

SPEAKER_00 36:17

Yeah, exactly. I mean, and you can think about it, right? So, how do we train a scientist? You go and watch someone else do it. Now imagine if the only way that you could train someone else in your lab is that you're writing a paragraph in paragraph form about what you did and you're hurrying up to do it because oftentimes that's not the point of your awesome manuscript that you're hoping to wow someone else with, right? So, so really PRISM was born out of this need for transmitting expertise. So I actually want to share a story about how we failed to do this before getting to Prism. And so one of the things we tried to do is we had built Cultivarium had built a portal. And so this portal indexes all the different organisms, their ID, the cross, you know, three to five different databases. And that was a pain in the butt because, you know, what is your ATCC ID versus the PubMed ID versus the RefSeq ID? These are the actual problems that people have. When we made this table, that was actually we thought we hit big with product market fit because scientists would talk to us and say, wow, I want that. I would just want that table. But we have built out this whole system and it's beautiful, it's still there. And people come in and use it as almost an encyclopedic type of thing where they go in, they check a thing. So we really love it having this resource for people. But the point that I wanted to make is we also made a space for what we call lab lore. And that was the core of the idea of you know, what is it that isn't made, it doesn't make it to methods written down in prose. Maybe you could write a little tip. And it's uh I guess another simpler way to put it, it was we were just trying to do Reddit remote microbiology, right? Now, having said that, scientists, what can you say? What we found from experience is that some scientists would write, but other scientists were very worried. They said, Well, if I ask something here, maybe that might be maybe maybe I look silly, maybe I just forgot a thing. And you know, we thought, well, we didn't think at all that someone would feel dissuaded from using this knowledge sharing tool. And so we thought, okay, maybe that's too much pressure, at least within the scientific ecosystem. Let's actually try a different way. So that's what brought us to prison, which I guess now it's been very long-winded. Let me describe what it actually is. This what we have done is you don't a pair of glasses that record audio and video. And purposefully, we did not choose meta glasses or any augmented reality or virtual reality glasses that others have tried. What we want is we want people to think about these glasses as if they're thinking about donning on a pair of gloves or a lab coat. It should be out of the way, not distracting. Let a scientist do their job. And what these glasses do is they will help you record what's going on, take verbal notes, and after you're done, that video and that audio gets transferred onto the Prism server, which then uses these foundation models to basically auto-document what has happened. Right. And the goal there would be that you don't have to think about all of the mental labor of writing everything down, at least for myself. This was a really funny thing that we run into. I'm a post-it guy. So I will the way I take notes is I scribble random things down on post-its, and that gets posted everywhere. So I look like I belong in some psych ward. Whereas my other scientist friends are very much a no, I write it down and it's in my, you know, my awesome trapper keeper full of college ruled paper. So so there's this huge diversity of how people take notes, and we just want it to standardize. So Prism now has been we have 30 plus labs all over the world. They're contributing videos that range anywhere from you know microbiology to chemistry. We have people doing just soldering and engineering about building devices on it. So we're

Prism And Capturing Lab Knowhow

SPEAKER_00 40:29

hoping that this will become something that scientists and engineers want to do to share their craft and to be able to distribute their expertise and have other people level up by watching and studying their videos.

SPEAKER_01 40:45

Very cool. And do you can are they citable? Like uh is there a way to incentivize people to participate beyond, you know, goodwill, we would hope, would be enough, but but incentive of credit is also nice.

SPEAKER_00 40:59

Yes, of course. So we have not added DOIs to it. We've been asked and we've thought a lot about it. It's on the roadmap. I don't know where it is on the meeting the threshold of actually getting time to work on it yet, but we thought about that. And then, you know, what's also interesting is we also thought we we took a quick look at other types of things that work like this. So, for example, Jove is the classic journal of visualized experiments. That's the more formal production grade. You want to do this protocol, we're coming, we're bringing in a bunch of production staff, and we're gonna do this thing. And then unfortunately, but maybe sensible sensibly it's paywalled, so it's hard to access, right? And you know, we thought, look, there's not that much site, there's not a lot of citation going on there. So we're just focusing on how do we get people to use it, distribute it, and yes, we are still working heavily on what is how can we better incentivize scientists to participate.

SPEAKER_01 41:57

Very cool. Well, so we've spent some time here talking about a variety of tools, enabling tools that you have made.

A Bio Cement Microbe Breakthrough

SPEAKER_01 42:05

Let's talk about some of the actual examples of weird organisms. Are there any favorites you've had or things that you've seen people do with your tools that excite you in light of either a particular application relevant to planetary health or this vision you have of being able to construct an ecology and choose the right organisms? Yeah, what have we seen? Yeah.

SPEAKER_00 42:31

So I get this question a lot, and almost without exception, I will reject this premise by not naming a single organism because again, I'm trying not to have this uh uh application space to one on one mapping to Latin name. But for you, Paul, and for this podcast, I will make an exception. Uh-huh. So I'm so grateful. So so and that is really just to say one of the key learnings that we've had at Cultivarium is that we want people to recognize the value of having tooling, of being able to get genetics to work in a variety of different strains. But, and this is without any negativity, most people will still identify with oh, you got you solved that particular organism. Now I now I'm paying attention, right? So what does that look like? So recently, I just picked the one off the top of my head, sporosorcina pastiuri. It is a microbe that performs biosementation. And so this is a strain that has been of great interest for microbial in microbial induced calcium precipitation. And so that helps you make cement like materials. And, you know, of course, cement, big, big greenhouse gas emitter, all of that. Are there ways in which we can use biologically derived cement? Maybe through a strain like this. I'm gonna Ignore for a second the extremely important aspects of technoeconomic analysis. So we're just gonna ignore that for a for a second. The key with this particular strain, sporososcina pasti, uh from DSM33, is that people have tried to get genetics to work for a very long time. You know, you talk to people who were working on this even up to 30 years, people have been trying to get this to work without getting genetics to work means being able to being able to get DNA through, being able to knock a gene out, being able to, you know, do some induction of the gene, turn it on. And so we thought, hey, let's give it a whirl. Let's see if we could crack this thing. And so we used all of our tools, we used different DNA delivery methods, we used our pools, our libraries of different promoters, origins of replication, antibiotic resistance, all that stuff. Now, what actually happened was kind of cool, which is that we we did a quick screen of this and we didn't succeed, it failed. We actually were not able to transform it. We thought, oh wow, this is really cool. We're gonna put it up on our pedestal of really hard problems to solve, but maybe later. But then what happened is we also underwent this natural isolation campaign where we just went out to the wild and just collected samples, got microbes, and then tried to transform those under the umbrella of we want to learn about how we learn. So we want to actually increase the breadth of what we're trying to transform. And so what actually happened was we were able to transform what was annotated computationally as a relative of sporosycina pasturade. And we thought that's sort of funny. There's this thing that we isolated from the environment that we could transform, and there's this thing in a strain bank where we can't transform what's going on there. And so what actually happened was we took some of the key tools that we had, the specific parts, the genetic parts, and the methods that had worked on this environmental isolate, and we just poured it on sporosarcinopasteurite. So what that means is instead of doing a screen, and so I'll sort of nuance this for everyone, where a screen by its very nature means that it is not, you're not trying to be a hundred percent exhaustive about what you're testing and how much signal you're getting from the assay, right? So you tolerate some amount of loss. And so what happened is when we were able to triangulate that these are the few parts that we have found to work in this environmental relative of sporosarcina pasteurae, then we went back and purposefully tested those in the strains that we in sporosarsina pastae, those just those particular molecular parts, and we got it to work. And so there's a couple of cool learnings from this, which is our classic way of doing microbiology meant that there was 30 years and who knows how many people trying to bang their heads against the wall on just transforming this thing in isolation. But if we had widened our aperture and tried a bunch of this other stuff, surely someone may have lucked into this, right? Or serendipitously done what we've done because it wasn't as if we did anything particularly sophisticated and we don't think we did anything particularly lucky on environmental isolation. And so this really, then again, convinced me that we need to widen our aperture, we need to think about cohorts of organisms, and that was a way that we could actually enable a bunch of different microbes to be genetically tractable. So that's that that's sort of one mega example or mega answer from your prompt on an organism that is cool, people are interested in it. We've had folks reach out to us to say, hey, can we get a copy of your transposon library? We've sent some out, we we've shared genetic tools with people. So we're really hoping that the field of MICP or microbial induced calcium precipitation will be able to take off with these now added tools on spores or pasteurize. But I hope people are using other species and other strains as well.

SPEAKER_01 48:17

Very cool story. And a beautiful example of this creative destruction that you were talking about earlier, right? Where you are trying something in a different way than the established method, and it turns out to work. I mean, many times you try a different way than the established method, it might not work. But the fact that you're trying many different things then enables you to access stuff that the established methods may not access. Um very cool.

Expanding Into Fungi And Archaea

SPEAKER_01 48:44

So let's now talk about the future for cultivarium. So the welcome program announced last year extends cultivarium's work specifically into fungi and archaea, which have biology that bacterial tooling doesn't cleanly translate to. And it sounds like a lot of what you've been doing has been for bacteria so far, or at least what we talked about today. And so I'm curious what you view as the big challenges for that and the big applications that could be unlocked. Okay, so we have talked about a lot about what cultivarium has done. Let's talk about where things are going. First, in terms of the some of the science you'll do, and also I'm curious to talk about like the structure of cultivarium because cultivarium is an FRO, and maybe the first FRO, one of the first FROs. So cool to hear about it from that perspective as well. And so the Welcome Trust program announced last year extends cultivarium's work specifically to fungi and archaea, where you know the biology of bacterial approaches doesn't really translate or doesn't translate completely. And most of what we talked about so far has been for bacteria. And so I'm curious what you view as the big challenges for domesticating archaea and fungi and the big applications that could be unlocked, and why you chose to go in this direction.

SPEAKER_00 50:06

Thanks for that question. It's uh it's really fun to really think about where cultivarium can have an influence and where we can solve some problems. So, yes, we are the first FRO, and we're certainly the first to s to get to our five-year plan. And I will just highlight that in the original proposal for cultivarium, we did include all sorts of uh what we call critters. So octopi, other types of critters, which have really interesting phenotypes that we wanted to establish genetics for, or at least more sophisticated genetics for. So this is a what you're seeing with cultivarium starting in bacteria, then moving to filamentous fungi and archaea, is really our slow but hopefully sure process of moving up the complexity of the tree of life. So the way we think about it is number one, we cannot possibly solve everyone's problems. And in fact, other scientists are best to solve their specific scientific problems. What we want to do is we want to give them an initial enabling set of tools so that they can move faster and spend less time banging their head against the wall on things that you would hopefully want to have as if you just started working on E. coli, right? So we think about our role as cultivarium to deblock a lot of these fundamental technologies or tool sets. Now, through our work, what we've found is that there are some general themes, and I've shared how we think about those themes. There are things like how do we think about it? It's husbandry, how do you grow something in the lab? And we presume that the organism you're interested in has some complex phenotype that you can't narrow down to one gene or one contiguous piece of DNA that you can just clone into something else. It is so cool that and its genetics is so unknown that you just need to work reproducibly in this organism. So if you want to do that, then you need to know how to do the husbandry, bring it into the lab, grow it. You need to be able to get recombinant DNA in, and then you need to be able to build genetic toolkits for it, right? So molecular toolkits of what are the different parts. So, and then we have this other column, which I alluded to with our portal and our prism, which is all of the knowledge around either the collecting the data, accessing the data, analyzing the data, so on and so forth. These are all themes that we can bring to accelerating biology throughout the Tree of Life. And so the Welcome Trust, really fortunate to have their support after uh future support, because we had thought about how we could take these themes that we prototyped using bacteria onto the world of filamentous fungi. So lots of awesome fungal researchers, lots of amazing discoveries. But the types of tools that we're bringing there are really trying to bring some of the what we'll call modernized bacterial approaches into the filamentous fungi world, right? These are things like for many filamentous fungi, there is no digital growth record of how these strains grow. And if you've seen them before, and this is what amazes me about some fungal biologists, they can just look at a plate and say, that's the strain, it's been growing for this many days, and that's a funky color, but that's okay because it's about to sporulate, right? And here I am as an outsider thinking, I don't know, that looks like it shouldn't be in my fridge. So there's a lot of this work where it's a nearly simple matter of, hey, let's build some infrastructure to digitize a bunch of this growth, literally take time-lapse images, and let's also screen some of these, the really popular fungi on a bunch of different growth media. And in fact, we have started collecting that data. We've found that certain filamentous fungi just look strikingly different from a morphological perspective on different growth media. And that may be of no surprise to fungal biologists that have spent decades training in decades old labs to do this, but being able to provide this type of resource allows others to join and to be able to participate in discoveries or participate in applications. So we really think about trying to enable people with these types of tools. So the future of cultivarum really is to take a look at this range of problems, which what we've trained ourselves to be good at is how do we deblock people with these tools to just get people into the lab and trying things, right? That's if there was a Maslow hierarchy of genetic tools or of biological tools, we want to solve what we consider to be the very base layer. And then everybody else can go and just run with all the cool things that they're about to build. So we are thinking about other types of uh biological systems. You know, we're really interested anywhere from protists to organelles to all sorts of different biological systems that are really just underserved because there's very few labs working on these things who are just barely able to keep husbandry going, right? Not to mention, hey, how do I think about doing CRISPR knockout? What does that actually mean? How do I get a molecular biologist to design the guide RNA for my genome, which may or may not have been sequenced at all? So, what we'd like to do is we'd like to continue this journey. We like to talk to labs, companies, foundations. If you're interested in a set of interesting organisms, we have shown that we can do this work in building this base layer so that others can explore and discover.

SPEAKER_01 56:04

Very cool.

Rapid Fire Takes And Closing

SPEAKER_01 56:06

So I want to move on now to our rapid fire questions at the end. These are questions that we ask to all of our guests, and we get some very interesting answers. And so we'll start with the first question: what is a single book, paper, piece of art, idea, artifact, whatever that blew your mind and shaped your development as a scientist?

SPEAKER_00 56:27

Without doubt, it's the Dyson sphere. I had never once growing up thought that someone could think about the fact that a civilization would build a structure that would completely cover the spherical presence of a star in order to harness all of its energy. With how AI is developing, I can see exactly how that's going to happen and when it's going to happen now. But it blew my mind as a kid that that we could be thinking at that scale.

SPEAKER_01 56:56

Okay, what is the best line of advice that a mentor gave you?

SPEAKER_00 57:01

So I will I'm sure I'll say this about George Church. I'm sure he didn't mean it as a piece of advice, but here's my very quick response. I had gone to George with a problem, and I had given him a couple of options of how I thought I could solve this problem. And I was waiting for him to give me a choose this over this response. And he basically spent his time telling me about the positives of every response. And that I took as a way to say, you know what? Maybe we just don't know and we just have to try them all. And so that that really has kept with me over my career as I think about, well, do I really know enough to actually make a decision, or should I really just try a bunch of different things?

SPEAKER_01 57:48

I love that. And that really speaks to the what seems like a kind of playful attitude that you take to a lot of your work.

SPEAKER_00 57:56

Yes, I hope so.

SPEAKER_01 57:58

Okay, here's what I'm excited for because Henry, I know you as a person who really delights in hot takes. And so I'm curious, can you give us a hot take? What's one view you hold related to climate and environmental biotechnology that you think some others in the field might disagree with?

SPEAKER_00 58:16

Do more, talk less. Don't be worried about runaway effects. And so I'll say a few more words there. And I think that homeworld has played a good role in things like garden grants to try to spark activity. I've I joined this community, I think, since day zero, because I would love to have an impact. I would love for our work to have some route to impacting the goals of what Homeworld is about. What I found is that it is very difficult to understand with any clarity what that looks like, because you get lots of opposing opinions and opposing views, and that by itself is okay. But it makes me get the sense that we are stuck on some, we just not moving in some direction, right? And so I think that yes, we should be concerned about all the different effects and the runaway effects and all that sort of stuff, but I think that we don't really know, and maybe back to this advice that I interpreted, I think we just gotta try some stuff. So that's what I would encourage.

SPEAKER_01 59:23

All right. And the last question what is one aspect of personal development that you think biotechnologists need to spend more time on?

SPEAKER_00 59:32

So I think that there's a growing awareness of this now, but I think that biotechnologists, at least coming out of the era of synthetic biology, would be it's always a good thing to do a little red teaming of your own ideas to think, is this really the best place for biology? A lot of times we can agree that there are some really powerful capabilities that biology has. But to be honest, most of human most of society likes really intensive processes. We like to be efficient. And so what that means is we want to have a lot in a short amount of time, and it's got to be cheap, and it's got to be really efficient on energy usage, not in terms of what biology does, which is, you know, you can do this sophisticated thing with very little power. It's I want to be really efficient with my time. I want X amount of kilograms of this stuff with, you know, in whatever, an hour. So I think that those are the types of things where we need to put biology, try to accentuate biology where it is best used in light of other existing technologies that are already very industrialized.

SPEAKER_01 1:00:44

Wise words. Henry, we have come to the end of our time. Uh, this has been a very fun conversation. I'm really grateful that you took the time to come here and talk to me today. I hope you have a good rest of your day.

SPEAKER_00 1:00:57

Yeah, thanks, Paul. It was really fun. And we'll hang out soon and hope to see you.

SPEAKER_01 1:01:02

Right on. All right, thanks, Henry. Thanks for tuning in. I hope this has been educational and inspirational for you as you navigate your own journey to bring the best of biology into planet scale solutions. I'll be back soon with another conversation. In the meantime, you can stay in touch with Homeworld on LinkedIn, X, or Blue Sky. Huge thanks to our producer Dave Clark, along with Paul Himmelstein, Kayla Sims Austin, and Cario Sarsoza for making these episodes possible. Until next time, I'm Paul Reginotto, and this is the Climate Clive Podcast.