302 | Chris Kempes on the Biophysics of Evolution

0:00:00.1 Sean Carroll: Hello everyone, welcome to The Mindscape podcast. I'm your host, Sean Carroll. Have you all heard about the centimeter long bacterium? I had not heard about this until this podcast, as you will discover, but just a couple of years ago, scientists found a kind of bacterium called Thiomargarita magnifica, which can grow up to a centimeter in length. It's a tiny, little tube-like thing, so it's not a spherical centimeter-long thing. But that's very creepy to me. I don't want any centimeter-long bacteria climbing around anywhere near where I am. And I bring this up because I did... Chris Kempes, today's guest, introduced me to the possibility during this podcast. I had never heard about it. But it shows sort of two sides of a certain coin for a bacterium that is about the size of a coin. One is, what we will be talking about in the podcast are the existence and usefulness of physical constraints on biological organisms and their evolution. So biological organisms are embedded in the physical world. They obey the laws of physics, and therefore they need to use good, old, sensible allowed physical mechanisms to survive, to metabolize, to eat, to move, all of these things.

0:01:17.1 SC: It should be unsurprising that the existence of the laws of physics will provide constraints on what kind of architectures and morphologies and sizes are allowed in the realm of living organisms. And of course, you can apply this idea, you can actually do it quite quantitatively and specifically to simple organisms like bacteria, and under some very reasonable assumptions, you can derive the smallest size of a bacterium possible and the large size of a bacterium possible. And these Thiomargarita guys are way larger than largest size possible. Of course, that turns out because they violate some of the assumptions you made. They're actually not like one big blob, they're almost like sausage links kinds of things glued together to make a long tube to grow to that centimeter size. But it's both an illustration, to me, of the power of physical constraints in biology, because you can't do anything. You got to obey the laws of physics, but also the cleverness of biology in figuring out ways around what you thought were physical constraints.

0:02:21.2 SC: So in today's conversation, we're talking to Chris Kempes, who is a faculty member at the Santa Fe Institute, a bio-physicist, I think that's safe to say, or a physical biologist, maybe. And the theme running through Chris's work is applying these physical constraints to life in all of its forms, so starting from viruses to bacteria, but we will also be talking about the transition from prokaryotes to eukaryotes, the existence of nuclei and other sub-structures in cells that are characteristic of eukaryotic life, and then the transition to multicellularity and even a little bit about larger macroscopic animals and their constraints.

0:03:02.2 SC: We live in a world governed by the laws of physics. Don't forget that. And those laws of physics will matter even for biology. Occasional reminder, here at Mindscape, we have a Patreon community that you are welcome to join. Just go to patreon.com/seanmcarroll, and you can leave... Oh, actually, we're changing how... We're changing the funding mechanism. I'm not changing it. Apple is forcing Patreon to do it. So anyway, there's gonna be a monthly subscription fee to Mindscape if you want to be a supporter, but it's not gonna be that much money. Actually, I got to do that updating now. And then you can be the kind of person who asks questions for the Ask Me Anything episodes, and you get ad-free versions of the podcast and a feeling of belonging that is useful to everyone in our troubled times. So with that, let's go.

[music]

0:04:06.5 SC: Chris Kempes, welcome to The Mindscape podcast.

0:04:08.8 Chris Kempes: Thanks, Sean. Really glad to be here.

0:04:10.7 SC: So you know, you're a faculty at the Santa Fe Institute, where I'm fractal faculty, so I guess my boundary area might be larger compared to my volume than you as just a regular faculty member since I'm fractal. But in both cases, we're supposed to be thinking about complexity. And it seems to me that even though SFI studies lots of different things, life, living organisms are kind of the paradigmatic example of a complex system, right?

0:04:41.8 CK: Absolutely. Yeah.

0:04:42.4 SC: Much of what I hear people talking about when we talk about complexity, clearly, they're either thinking about life or inspired by living beings. Is that a fair kind of thing?

0:04:51.4 CK: I think that's incredibly fair. People often say complex adaptive systems to define complex systems or complexity, and so the adaptive part really matters to many people. And so life is the quintessential adaptive system. And what we mean by that, typically, is that for an organism just to survive, just to persist, it has to deal with this constantly-changing set of environments, it has to respond to all sorts of new predators, new co-operators, new ecologies, and to do that, it has to constantly mutate and evolve and adapt. And so that's one way to really define complexity, which is systems that have to continually find new solutions to an ever-changing world.

0:05:36.1 SC: It's an interesting perspective, because, of course, there's many ways of defining this. I kind of don't care what the definition is, obviously. What matters is what the things do. But that's a very kind of interaction-ist perspective you just gave. It's about how the system adapts to the world rather than a completely internal definition in terms of hierarchy and organization and parts and holes and things.

0:06:00.9 CK: Yeah. I think that's a great point. I think for me, I really focused on that hierarchy and internal structure and all the ways that organisms come to find solutions. To me, that's all a consequence of needing to adopt. I like to say if there's one environment out there with only one limiting resource, and it never changes, you only need one very simple organism to survive in that. And in fact, it has no benefit in becoming more complex, in gaining higher levels of structure and developing all these wonderful fractals on the inside like we do with our vascular systems. And so if you have that really boring environment, you don't need adaptation. You just need to find one solution and stay there forever. In fact, all mutations are bad. And for us, some mutations are good. And so I think that is why I would sort of say, you need that change, you need some external motivation to get all the richness that we see in living organisms.

0:07:08.9 SC: Good. Okay. So to get down to business, I noticed that you opened a paper you wrote a few years ago with the provocative statement that, "Organisms are subject to the laws of physics." These are bold words, Chris. Does everyone agree with this? Is this gonna be a major constraint?

[laughter]

0:07:29.4 CK: Yeah. Well, it's interesting, 'cause it's a blatantly obvious statement, as you're alluding to. And yet, it's something we take for granted. We were handed, in the modern world, this rich set of organisms, this vast diversity that we see in all these different ecosystems that we go into, amazing creatures that have toxins and weird ways of moving around and all sorts of different intelligences, and so we look at that world, and we start to... And biology has classically started to try to tell stories about those organisms or explanations focused on the richness of their features. And I think we often forget this very simple fact, which is, those organisms are situated in a physical environment. They're situated in the laws of physics. They're situated in the laws of chemistry. And so a lot of what I do is to try and use the other end of the spectrum of thinking to say, if all we focused on was the fact that there are these physical laws that organisms live in, how much can we say about biology? How much of all of that richness and amazing variation can we explain from just focusing on the laws of physics or laws of Chemistry? That's not every project. There are some places where you can't say very much about organisms from that perspective, but a lot of the game we play is to move into that space and say, "How much can we explain from this simple perspective?"

0:08:52.3 SC: I'm unsurprisingly going to be sympathetic to this perspective, but I guess, I can see why one might not bother with that perspective. The variety of living organisms is so incredibly large, when you think about what evolution has accomplished with all the species, et cetera. It seems you could do anything. And here you are saying, but yes, the center of mass of the organism will follow Newton's Laws of Gravitation or something like that. And I guess a typical Biologist might say, "So what? That's not much of a constraint."

0:09:23.1 CK: Exactly. And so I think you could say, well, imagine I'm in some sort of game theory situation with another organism, where it's trying to eat me and I'm to build up these defenses. So let's say I keep building this longer and longer spine on the outside of a very hard shell, and I do that so that it's harder and more painful for me to fit in something else's mouth. This is the story of different phases of evolution in the history of life. Now, if I keep doing that, though, eventually, I start to get a spine that is so long that the leverage on it means it's really easy to break. And so either I can't grow the spine any longer, or I have to make it much wider, or I have to use a different material to make that spine.

0:10:09.6 CK: So this is a place where you'd say, well, the evolutionary dynamic is really about this predator-prey evolutionary chase, where one's trying to eat the other and the other is trying to build up defenses, but at some point, and probably at every point, those decisions run up against the laws of physics. So to build a great defense, which you'd say defense is only in the context of a predator, but still to build up a great defense, I have to do so in a way that is also optimizing certain physical constraints, and that's the key. And so it's that every time I'm making an evolutionary consideration, I'm also making a physical one. And actually, Galileo was the first person to point this out, as he did with his drawings of bones and explaining how, as organisms got bigger, their bones had to get much wider just so that they wouldn't break. And so I think this is something, again, we often take for granted.

0:11:02.0 SC: And the good news here is that by the laws of physics, we really don't need to worry that much about quantum field theory, et cetera. It's more or less classical Newtonian stuff.

0:11:11.5 CK: Well, there is interesting quantum mechanics in systems like photosynthesis. I don't work so much on those. I'm mostly concerned with packages of biological stuff. So I sort of think the viruses are sort of as small as I get, so I don't go quite to the quantum mechanical scale. But yeah, you can use a lot of classical physics to describe much of what we see in organisms. Now, where that gets complicated is where those classical physics interface with complicated, say, mathematical geometries. So a lot of the work that's been done at the Santa Fe Institute, Geoffrey West and Brian Enquist and Jim Brown did this work in the late '90s, was to show that when you start to think about optimizing, even classical physical constraints, the best way to do that is with a fractal geometry. And so then, the mathematics of that geometry become more complicated, but the constraints you're optimizing are still classical physics constraints.

0:12:09.1 SC: And just so... I think, maybe people might get this mixup a little bit. The constraints we're talking about are not necessarily external constraints, like being fit into a box or something like that. There's sort of constraints on what you can do. I don't know. I always start in these podcasts by making statements and then saying, "Is that right?" I should phrase them as questions. Should we think about being fit into a box, or do you mean something slightly different by the words "physical constraints?"

0:12:41.7 CK: Yeah. And what I mean by physical constraint is, really, there's a physical law. And as you know well, physical laws often apply at certain length scales or certain amounts of mass or that sort of thing. But over some range of sizes, there are certain physical laws that are present, and those informs how all the physics at that scale operates. And that's what I mean by constraints. So gravity is really... It's a law of physics, but we could think about that also as a constraint in the sense that if I get taller, I have to deal with different forces of breaking that O2 gravitational forces on our planet. So it's really that sort of thing. Now, you can have more abstract constraints, some of which emerge in time, over evolutionary time. So as organisms evolve and become more complicated, you can get new sorts of constraints. For example, I already put the brain here, and so if I'm gonna put an eyeball somewhere, where should I put it? And that's a constraint that's related to an evolutionary decision that was made in the past, and we often call those emerging constraints or physiological constraints. But often when I'm saying physical constraints, I just mean the laws of physics; diffusion, gravity, fluid flow, electricity and magnetism. These are the sorts of things that we're thinking about.

0:14:11.8 SC: And is this pointing in the direction of explaining examples of convergent evolution, like dolphins look more or less like fish, because given what they're trying to do and the constraints of laws of physics, that's what an inefficiently-swimming thing is gonna look like?

0:14:26.2 CK: Yeah. That's a great way to put it. And I often like to say that evolution according to physical constraints is sort of the ultimate convergence. It's the thing you can most rely on. It's the thing most likely to happen. And we wrote a paper where we said, you'll get optimization of traits if those traits depend on a very dominant physical constraint. And what I mean is, to your point about quantum mechanics, quantum mechanics certainly apply at the scale of entire trees. They're not a dominant constraint. So many of the quantum mechanical effects have been averaged out of all of these different particles that I don't really need to use, say, the Heisenberg's uncertainty principle to understand what's happening for a vascular plant or a large tree. However, that constraint is still there. But gravity is a dominant constraint. And so if traits are connected to a dominant physical constraint, evolution will be able to see that, and we'll get traits that look optimized according to those physics.

0:15:29.9 CK: And then those traits had to be independent enough of other traits. So there are many cases in biology where you would like to optimize something, but you can't for other reasons. So a great example is our optic nerve that connects our eyeball to our brain is longer than it should be. And many people have talked about, "This optic nerve seems much longer than you would want in some optimal case." And part of that is likely due to the fact that you had brain structure and you develop this eye, and then it's really hard to remodel the brain. It's really hard to move around different modules of the brain, and so you deal with a slightly longer optical nerve because it's just too hard to change everything else. And so that's what we mean by traits have to be independent enough of other traits in order for these physical constraints to be something that evolution can see.

0:16:17.3 SC: That might be a good example of... I had a recent podcast with our mutual friend, Brandon Ogbunu, and we were talking about fitness landscapes and seascapes and so forth. The fitness landscape changes over time, so at one point, the direction to go in was to build a brain, and only later, that you need to make an optic nerve, and so you're stuck, kind of, in a little valley where only certain things are possible.

0:16:42.8 CK: Exactly. Yeah, exactly. And so what we often... In that same language, what we would say about some of these physical constraints is they create these huge valleys that are impossible to miss.

0:16:52.9 SC: Yeah. Good.

0:16:53.9 CK: And so even though there's roughness on the surface, it's really like a gently undulating grassy hill that is all part of one big valley, one big watershed where everything rolls to the bottom, and that's just because the physics is so strong. The physical constraints are so strong. You can't help but wind up in the bottom of the valley. But there could be emerging constraints that create local valleys that you can't get out of, as you were mentioning, and so forth.

0:17:23.5 SC: While we're at this very high level of abstraction, we'll get our hands dirtier in just a second, but, the other thing that I liked much in one of the papers that I've read of yours, the idea of competing constraints. There's one constraint that makes you wanna grow bigger, another one that makes you wanna stay small, and you can almost quantitatively predict what is the smallest kind of organism you could get, what is the largest kind, what is the optimum size.

0:17:50.2 CK: Exactly. And there's a lot to say about that. So one really interesting thing is that, often, the only way to explain what you see an organism structure or function is from a consideration of multiple constraints. And so what you do is you say, we find some way to write down what in some fields would be called a global cost function, where you put all of the terms of different constraints in that same cost function, and maybe one trait interfaces with multiple different constraints, and then we just optimize over the whole thing and find some global optimum under multiple constraints. This is an old idea and engineering, and that often is the best way to explain complicated structures. So for vascular plants, which I was mentioning, the original work on that showed it wasn't just about gravity, it's also about fluid flow through the vessels of the plant, and is about trying to fill space with all the end points of the network so that you can either distribute leaves and space to... Atmospheric space to uptake sunlight or to distribute cells in the body and feed them all with this vascular network. So that's a place where you sort of have three main constraints that you're optimizing over.

0:19:02.7 SC: This is more like an end of the podcast kind of question, but does this mean that if we discover advanced interstellar life, alien life, that we shouldn't be too surprised if there's some commonality of forms in the morphology?

0:19:19.2 CK: Absolutely. So, I think we wouldn't... I'm willing to make bets about that actually.

0:19:24.0 SC: Okay, good.

0:19:26.4 CK: That if you get large multicellular organisms, they will have fractal-like vascular networks with a certain very specific fractal structure that have been outlined in the whole thread of work that I was discussing before. And so we have some evidence for that in that. The vascular system in plants evolved independently of the vascular system in mammals, and yet they share a huge amount of the same structures. So that's convergent evolution with a huge evolutionary divergence in time. One organism is motile, the other organism isn't. One uses sunlight for energy, the other eats things that use sunlight for energy. One's warm blooded, one's not. And yet the vascular systems we find share certain commonalities because they are roughly the same size in obeying the same physics.

0:20:18.0 SC: Good. Okay. Now, let's see...

0:20:19.9 CK: Yeah. I think these are sort of the ultimate astrobiological convergence.

0:20:24.6 SC: It makes the Star Trek budget help under control. All you need is to slap some prostheses on a human being, and you get an alien. Physicists of course like to look at toy models and spherical cows and so forth to put their theories to the test. So let's think about bacteria. They seem to be simple unicellular organisms that we can really do some math on, perhaps. Well, tell the audience about bacteria. One amazing thing is the range of sizes that we get in bacteria. This always knocks my socks off.

0:21:00.6 CK: Yeah. So bacteria, we think of them as these tiny sacks of stuff, which in some ways they are, but even, as Sean was just saying, even then, we get a factor of 10,000 in cell size. So that's four orders of magnitude from the smallest cell to the largest cell, and the smallest cell and the largest cell don't look like each other in lots of different ways that we can get into. And so, people often just say bacteria as the small stuff, not realizing that there's an entire world in bacteria of diversity and different cell sizes and trade-offs and lifestyles and environments. And so even there, we're talking about half the range of body size differences that we see in mammals in terms of orders of magnitude. And so a factor of 10,000 is definitely nothing to sneeze at.

0:21:54.3 SC: And the, just to get, again, our non-biologically literate audience up to speed, the bacteria don't have nuclei, they're eukaryotes and they're friends with the archaea, which are the other kinds of prokaryotes, right?

0:22:13.2 CK: Exactly. Yeah. So bacteria prokaryotes, they have some amount of internal structure, and this is something we've learned a lot more about recently, but in general, you can think of them as some sort of membrane, and sometimes that's two layers of membrane. And then inside that membrane is a bunch of more or less free floating stuff. That includes a free floating genome that's not packaged in a nucleus. It includes a bunch of large molecular machines often called macromolecules, all of the enzymes and proteins of the cell. And mostly the stuff just diffuses around and interacts in a very complicated but messy network to do all of the functions that a cell needs to do to metabolize, to synthesize new parts and eventually to divide, to import and export material from the cell, and for larger cells to swim around and follow nutrients and make some decisions about their environment. But yes, they lack a lot of the internal structure that we see, even for other unicellular organisms. So there are other single cell organisms that have more internal structure. Bacteria tend to be these very simple things, not quite the simplest life because I'm a, viruses are life too, sort of person.

0:23:33.5 SC: Yeah. Very controversial.

0:23:34.2 CK: And we could get into why I think that's the case. So I would count viruses as the smallest life and the simplest example of really an enclosure and some genetic material inside that enclosure. Bacteria then have this active goopy stuff inside.

0:23:53.4 SC: But the DNA, they're just like wandering around, like bouncing back and forth inside the bacterium.

0:23:56.8 CK: Yes. Exactly. The DNA is just... So it's actually a circular chromosome in many cells. So it's a DNA that forms this entire loop, and that loop is not a rigid loop, it's malleable and moving in time. And so think about sort of a circular chain inside a cell that's jostling around and sort of moving all over the place while interacting with a bunch of molecules that read off the genetic information and copy the genetic information when the cell divides. And so it's in some ways sort of amazing that without much structure in such a messy environment, we get such reliable organisms. There's a sort of unbelievable number of prokaryotes on the planet and they evolve very quickly. They live in every environment you can imagine. They divide reliably. They're just quite amazing creatures.

0:24:53.8 SC: And so, despite the fact that four orders magnitude in the size, that is to say a factor of 10 to the four, that's pretty big. It's a large range, but there is a lower limit and an upper limit. And this is where you're gonna earn your keep, telling us why the physical constraints help us understand those limits.

0:25:11.5 CK: Yeah. So something we like to say is that, if you have a dominant physical constraint, it will also tell you where there's a wall, where something becomes as methodically limiting. And what I mean by as methodically is just something goes off to infinity, and you can't keep up with it, or something goes to zero, and then it does you no good. And so often, if you're writing down some optimization equation, and it has physical constraints in it, you should expect to see a limit at some scale. And so we see these in lots of different groups of organisms, but particularly in bacteria, we have pretty well worked out what the lower limit and the upper limit are. And so if you think about the smallest bacteria, they have this membrane that is shrinking down, if you imagine a spherical cow, and here we really can't imagine a spherical cow because the small bacteria...

0:26:03.9 SC: Bacterium.

0:26:04.0 CK: Are little spheres.

0:26:06.9 SC: Yeah.

[laughter]

0:26:07.4 CK: So imagine the small spherical bacterium. And as it shrinks down, the membrane has its own finite thickness. And so imagine you have a really thick tire and you shrink down that tire, eventually the rubber is touching the rubber on all sides and you don't have a tire anymore with something in the middle. You have just a hockey puck. The same thing is happening for cells. So they have this membrane that is very thin until you get to really small sizes, and then it starts to represent a large fraction of the cell size. And then you still have to inside of that, in the remaining volume, fit all of the stuff. So you have to put the DNA in there. You have to put a few functional proteins to actually run the metabolism, and you have to put in this amazing molecular device known as the ribosome, which is sort of this generic tape reading device that takes information from the DNA, and turns it into functional proteins.

0:27:08.6 CK: And so in these small cells, the DNA starts to take up roughly half the cell volume. And so you're running into all of these space constraints, where you almost can't fit in just the information of what the cell is. You think about how strange that is. The storage system for information becomes sort of half the cell volume. And you get to a point where every single gene that you would eliminate kills the cell. So every gene knockout, as we call it, is fatal. And so you can't eliminate any more genes. You're stuck with this minimal genome. And so the smallest cell is defined by this point where you have a minimal genome, a handful of ribosomes, and the functional proteins that do all the other metabolic and physiological aspects of the cell.

0:27:55.5 CK: What's amazing is that we write down energetic models of the cell, and we think about what's the energy budget of the cell, what it has for running biosynthesis, the production of new stuff, and what it has for maintenance, which is just repairing and the repair and upkeep of existing stuff. Then at these tiny cell sizes, the maintenance starts to take over the entire metabolism. So you start to look more and more like a cell that can only repair itself. It actually can barely grow any new stuff for replication. And this space constraint and this maintenance constraint where maintenance metabolism becomes the whole metabolism, both happened at basically the same tiny cell size. And so we think it's this dual constraint that sets the smallest possible bacteria. Amazingly, those predictions have agreed with all of the world record holders for smallest bacteria and so we... And it's where you transition to seeing tiny viruses that crop up at these really small cell sizes. So we think this physical constraint tells you the smallest possible bacterium.

0:29:06.1 SC: Knowing that our audience is raised on watching movies like Antman and so forth, we have to say like, why can't you just make the cell wall smaller?

0:29:16.3 CK: And you could, and that's a really great question, Sean, 'cause that's a game, especially in an astrobiological context, we like to think about a lot, which is, well, but the DNA takes up a certain volume and we're committed to a certain chemistry there. So what if you change that a little bit? And the membrane is made up of this lipid bilayer, and what if we used a smaller molecule than lipids? And so one can really absolutely start to play that. And this is a place actually where we're separating the contingencies of evolution from what might be truly universal. So yeah, if you could find a smaller enclosure that does all the same things that our membrane does in terms of it's flexible enough, it keeps the inside separated from the outside. It has the right sort of hydrophobicity, all those sorts of things.

0:30:06.8 CK: Yeah, there's certainly molecules that would be thinner in a layer, and then yeah, you probably could get a little bit smaller. So I think this is a case where we think we understand the constraints that matter, why the smallest cells on earth are the smallest they can be. And then we could play a little bit of games with chemistry to say, how much could you push that lower limit in size? In fact, that game was played a bit because the Allan Hills meteorite, which was this meteorite from Mars, where under a microscope people thought they were seeing structures that really looked like cells or the remnants of cells or some sort of precipitate from a cell. And this is famous in the astrobiology community because Bill Clinton gave a press conference saying we might have potentially found life on Mars or evidence of life on Mars.

0:31:00.1 SC: I think infamous is the word you're looking for, not famous.

[laughter]

0:31:04.2 CK: You're right. I forgot my prefix. But this, because of how big a news story this was, and the implications, the National Academies put together one of these national research council reports, where they gathered experts from paleontology and biophysics and biochemistry to try and say, are these things too small? And one of the initial sort of critiques was, yeah, but these structures are really small. They're smaller than the smallest cells we see. Are we actually talking about fossils here? And so in this report, which is really wonderful, people work through all these different angles to say, yeah, this is probably too small. So people work through, okay, what's the most curvature you could get from some polymer or some sort of molecule? And are these things below the curvature radius for that molecule such that we don't think you could actually encapsulate with molecules at this scale? And then people, and then there's all sorts of other rich material in that report. And basically every paper in this report says these are too small. But this is exactly the game that we have to play to try and understand what we might be looking at somewhere else.

0:32:19.2 SC: But aside from the possibility of novel molecular structures we are running up against, I'm gonna have to take back what I said earlier. We're running up against quantum mechanics, 'cause it's the size of the atoms that ultimately sets what we can do here. Right?

0:32:35.7 CK: That's a great point. Yeah. You could. You could even play a... I don't know if we have an astrobiology for other universes yet, but you could play that game. You could say imagine you were in a universe where the atoms were smaller.

0:32:51.8 SC: Change the mass of the electron. Yeah.

0:32:53.0 CK: Yeah, exactly. Assuming all the other physic... You can still get all the other physics you need, what's the smallest possible cell in such a universe? And I think that's... Yeah.

0:33:00.8 SC: Yeah. People have done that in the context of the anthropic principle, maybe not exactly that, but they've definitely asked whether you could have the chemistry.

0:33:05.7 CK: Right. Exactly.

0:33:08.5 SC: Okay. I guess that makes perfect sense to get an organism at all, an organism that has its own DNA and could be produced by itself. There's probably a minimal size you can get away with. There's certain ingredients you can't do without, but why is there a maximum size to bacteria?

0:33:27.6 CK: Yeah. So the maximum size is really interesting. That actually concerned us for a long time. So when we first sort of started uncovering what we think was limiting the smallest bacteria, we then had nice sort of laws for how things like growth rate would change with increasing cell volume. And so we thought we understood the trends in bacteria as well as the small limit, but then we had this huge puzzle about, why is it that you don't just keep being bigger and bigger bacteria? So what's happening, the biggest bacteria are growing really fast per unit volume. In fact, growth per unit volume is increasing as bacteria get bigger. You have lots of metabolic power. Everything seems great for the biggest bacteria. Why do you have this major evolutionary transition to unicellular eukaryotes which we can get into in a bit, where everything shifts and growth rate decreases and all of that.

0:34:24.9 CK: Why is there a motivation for that? Why not just, why aren't we all just giant sacks of bacterial stuff walking around and building cities and talking on podcasts? And what we proposed as a hypothesis in one of those early papers was, it's likely that as growth rate gets faster and faster, eventually, the basic biochemistry just can't keep up. So if I have some chemical reaction that can only go at some maximum rate, or I have some molecular device that can only work at some maximum rate, eventually it can't keep up with my overall growth rates. So it just can't keep up with how fast I'm trying to move. I could pump more and more energy into it, and it just can't turn over quickly enough. And we are very used to thinking about that with macroscopic machines where they all have an upper threshold where you're just trying to push them too quickly.

0:35:18.8 CK: And so we said, let's look at all of now the detailed biochemistry, all of the macromolecules, all of these different components of the cell, and ask how they change as cells scale up from very small sizes to very large sizes. And as we do that, we see uh-oh, there's this one component, the ribosome, which again, is this device that's sort of the center of biosynthesis. It's turning genetic information into functional proteins. It's actually building the functional proteins. And we realized for it to keep up with this runaway growth rate, eventually you would need more of that device than you have cell. So you would need to pack more... You need more of this machine that can actually fit inside the cell. We call this the ribosome catastrophe, and it sets the largest possible cell size.

0:36:08.0 CK: What's really happening there is that the ribosome is making more ribosome, so it actually makes some of its own components. And so you reach, it has its own replication rate. So it's a machine that replicates itself with one rate, and eventually the cell growth rate starts to push up against that ribosome replication rate and that is an impossibility. And so that's where you get one of these, something goes off to infinity. You need infinitely many ribosomes, that can't happen. You overpack the cell and so we think there's this upper size limit there.

0:36:40.9 SC: Well, this sounds like something that you could calculate mathematically and then as scientists, you're gonna go compare with the data.

0:36:47.8 CK: Exactly. And so we did that. So we said, okay, we know where this asymptote is, and let's look at the data. And so interestingly, what you see is that you get less and less species, less and less known species of bacteria as you get closer and closer to this wall. And then we said, well, what's the world record holder bacteria? And I wanna be clear here that I'm not talking about these recent world record holders for the longest bacteria. So there are these almost centimeter long bacteria.

0:37:16.4 SC: Oh my God. I'm gonna lose sleep now.

0:37:20.7 CK: Yeah, exactly. But the key thing to note about those is they're actually filaments. And so they're sort of like a colony of many little rod-shaped bacteria, like E. Coli stuck end to end. And so why they form these very long filaments, there's many copies of the DNA, and their dimension in the other radial dimension is like an E. Coli. It's a small bacterium. So they get very long, but they don't get volumetrically big. And that's really what we're talking about when we say size in this case. So anyway, we look for the world record holder, volumetric organism, something that's roughly a sphere. And there we find an organism that's about a 100 times bigger than our upper bound.

0:38:04.9 SC: Okay.

0:38:05.5 CK: So we thought, okay, this is interesting. How is this thing a hundred times bigger than our prediction, especially since we'd been so pleased by our lower bound predictions? And there's two things to note. One is, this organism has this huge internal storage vacuoles, so it has these big internal membranes that it uses to just store inorganic stuff, likely to be able to grow fast later. So it sort of hoards rare stuff and then uses it later. If you remove all of that and you say let's just talk about the active, sort of metabolic portion of the cell, you're still bigger than our limit. Not by as much, but still a little bit bigger. And here then the other crucial thing to note is that this organism lives in a really resource poor environment.

0:38:51.7 CK: So it lives in these deep ocean sediments running a metabolism that we know is a really poor metabolism, it just doesn't give you much energy. And so they grow very, very slowly. In fact, they're really hard to study and grow in the lab, and you have to do so in environmental settings. And so what that tells us is that this upper bound that we were describing really was a speed limit. We said, there's a fastest growth rate, there's a fastest replication rate. If you wanna get bigger than this wall, you just have to gross more slowly. And so that's actually what the next branch of life the Unicellular eukaryotes does, is they grow more slowly. And that's also what large bacteria that break this rule do is they just stay away from that fast growth limit, that speed limit.

0:39:40.4 SC: It's very much, I love that example of how science works. Like you have a theory compare it with the data you realize, I could have had a better theory or a different theory all along, but I'm learning something by looking at what nature actually does.

0:39:53.1 CK: Exactly. Yeah. And, it made us realize, to your point, exactly what we were talking about in our theory. What our assumptions were is it was about maximum growth rate. And so we're confident in maximum growth rate, but there's ways to break that law. You can just dial down your metabolism and be a slow growing organism.

0:40:09.2 SC: Is there any relationship between, Jeremy England's work on the statistical mechanics of self-replication? I've had him on the podcast years ago.

0:40:18.2 CK: Yeah. So, there's lots of interesting ideas there. And I think where it interfaces what we're doing is just to ask questions about what are the fundamental bounds on how efficient you can be, how much energy you need to perform a process and so forth. And we've done a little bit of statistical mechanics on the energetics of different cellular machines. So a bunch of us, including David Wilpert a few years ago, worked out sort of what we thought, the optimal efficiency of something like the ribosome would be and then compared it to the efficiency of the ribosome that we have. And so if you abstract the ribosome, what it really is, is a device for writing strings, for taking a random set of letters that are unorganized and writing those into exactly one string.

0:41:08.4 CK: That string folds up into a functional protein. And so we could compare that thermodynamically to any device that writes strings, and we find, that gives us an ultimate efficiency, the best that any device in the universe could do at this abstract process. And we found that the ribosome is at most 20 times worse than that limit. And I say 20 times off the limit that seems really inefficient. But for reference, our computers as we talk now, are writing strings at a hundred million times worse than the limit. So the ribosome compared to our computers is like many, many, many orders of magnitude more efficient than our computers at this string writing process.

0:41:49.9 SC: It's bumping up against the limit in some real sense, even though it's 20...

0:41:53.1 CK: Exactly.

0:41:53.9 SC: Factor of 20 way. Yeah.

0:41:55.3 CK: Exactly. It's feeling the ultimate physical limit.

0:41:58.2 SC: Interesting. Okay, good. But this doesn't sound like we're done yet. It doesn't sound like this field is completely, Is dotted and Ts crossed?

0:42:09.0 CK: No, definitely not. I think the statistical mechanics of cells is something people are working on in lots of interesting ways. As I mentioned earlier, I said, bacteria sort of mostly don't have structure. There's been this whole revolution really in the last 15 years around phase separation in cells, in bacteria and realizing that different types of stuff are actually forming separated phases. Think liquid in water and how you could have a little bit of, sorry, think about oil in water and how you could have a little oil droplet floating around in water. They're both liquids, but they're sort of separated in an interesting way. And that's happening inside of cells, and that has chemical consequences, that has structural consequences. And so I think there's a lot to uncover there. There's a lot of really fundamental statistical mechanics and thermodynamics to do within cells.

0:43:00.9 CK: And then I should say that recently we've worked out other limits at the large end of bacteria where we have come to realize how fundamental constraints of diffusion start to set an upper bound for bacteria where you need to add internal structure just to make things move around more quickly. And so there's a sort of diffusive boundary these start to run into, and then you have to add transport structures to sort of get around that. And that occurs also roughly at the same size that you're seeing this ribosome catastrophe. So just like the small end of bacteria, we're seeing a large end where multiple constraints seem to be limiting cells at pretty much exactly the same size.

0:43:40.9 SC: Well, this is good because it's definitely nudging us in a certain direction that other overly sized bacterium you talked about before, you mentioned that it has these storage vacuoles inside it. It seems as if when you get big, it begins to become helpful to become complex at least have more structure inside you than the ordinary bacterium does. And of course, we know famously eventually we're gonna make eukaryotes, we're gonna make cells by marrying or, I don't know, shotgun marriage of the archaea and the bacteria, to get cells with nuclei in them. So is that something, that seems like an out of context problem. It's like you're changing all the rules once you make a eukaryote because there's a transition to a whole nother kind of thing.

0:44:29.3 CK: Yeah, it's a great point. And I think what's interesting is that we would say once you get one of these walls, one of these sort of hard limits coming from physical constraints, it tells you you need to shift something about your architecture to get over that constraint if you're gonna get bigger. Now, there's lots of cases where you can imagine evolution never discovers anything that helps you get bigger. But if you see something that gets bigger, it certainly has found some new architecture. And so, one of the great challenges of modern biology is thinking through, how and why we got eukaryotes.

0:45:07.5 CK: And so eukaryotes, are, as Sean said, a byproduct of what we call an endosymbiotic event where one cell started living inside the other cell. And so you, in this case have a bacteria likely living inside of an archaea. And the bacteria inside the archaea becomes this metabolic structure known as as the mitochondria, which many people sort of associate with how life gets big and why life gets big, and how organisms like us can produce so much energy. There's lots of other reasons that you might want a mitochondria. It distributes many copies of the metabolic genome throughout the cell. Nick Lane has written really nicely about that. It distributes, certain sorts of metabolic processes throughout the cell in a structured way.

0:46:03.4 CK: All of this is sort of really beneficial. And we understand that it must be getting over limits, that we saw out of bacteria. I think our perspective would be the main limits that you start to see are about transport. And so I would almost say you likely need to solve transport before you solve these other packaging problems around different sorts of genomes and having lots of internal metabolism and so forth. I think it's actually, you just can't get substrates, the stuff that you use, products, the wastes, in and out of the cell fast enough with diffusion alone for large bacteria to make the metabolism function. So you have to start adding active transport inside the cell, ways to stir the cytoplasm, ways to move packages around in the cytoplasm just to overcome the sort of challenges of diffusion. And I think you probably need to do that before you start to get things like the mitochondria and other sort of endosymbiotic events.

0:47:05.9 SC: This is good because it's not like teleological, it's not like something is making it more complex. It's that the cell given what it wants to do is being faced with a problem and extra layers of structure are going to help it solve that problem.

0:47:19.3 CK: Exactly. And you could think of it as a drift process or a random walk process in evolution. And so if there is one of these walls, I think you expect an evolutionary time for a very long period of time things just keep hitting that wall and be stuck in a certain size range. And then you're waiting for a chance event where some fairly big change to architecture happens for one reason or another, and that gets you just on the other side of the wall and then you're off to the races in this other evolutionary trajectory. Then all of the physics of that region become a target for evolution. And then you may need to solve a bunch of problems, successively. But many people would say it was harder to get the eukaryote than it was to get cellular life in the first place.

0:48:08.2 SC: Took longer.

0:48:09.5 CK: It took longer. And I find that really interesting and surprising. And I also don't necessarily agree with that 'cause I think there's other factors involved. But that sort of thinking I think is really important to say, you can just stall out at one level of complexity for a long time. Nothing wants you to get more complex. It's just if you stumble upon the right solution that gets you over that hump, then you're into this new valley with no competitors, you've got a new niche and we know it happens there. Things diversify and go in every direction and you get lots of new solutions. But you have to wait for that chance of events to get you over the hump still.

0:48:49.2 SC: And this might actually help push me in the direction of greater understanding towards something that's puzzled me ever since I realized that even though the bacteria got embedded in the archaea to make eukaryotes and eukaryotes have nuclei, it's not the bacteria that are the nuclei, the nuclei are a whole separate thing. But if it's because it's pressing up against some limit at its current level of complexity and it needs to become more complex, maybe it's not surprising that it's becoming more complex in multiple ways. It's getting a nucleus and it's getting mitochondria.

0:49:24.9 CK: Exactly. And I think one of the reasons for that is once you move into this new space, you see a bunch of new challenges that you never had to deal with before. And likely each of those challenges needs its own solution. And so it's not then surprising that you get a bunch of complexity all at once. And maybe you don't get it all at once evolutionarily, looking back from our current vantage of all of these well-adapted organisms a billion years later, we would say, oh wow, all of these, there's many innovations. They're all really optimized and they all go together. Now, one of the biggest debates in how eukaryotes came to be is the ordering of events.

0:50:06.7 CK: It's countless papers written about what was first. It's even gotten to the point where we say, well, is that a mitochondria first argument? Is that a phagocytosis first argument? And so people really are trying to work through the ordering and have big debates about the ordering. And we don't have any fossils to tell us because these things are goopy and don't preserve well. And so whereas that debate gets resolved in dinosaurs about what happened first, it doesn't get resolved for the eukaryotes yet.

0:50:38.0 SC: And we're still in this realm of what we say eukaryotes. They're much more complex in some way than the prokaryotes, but they're still unicellular organisms, so we can compare them. One of the things you did was say, what kind of scaling remains true as you go from prokaryotes to eukaryotes, and then what things kind of break down and become different.

0:51:02.3 CK: Exactly. Yeah. So one thing that we found was, the way in which protein concentration scales with cell volume is preserved from bacteria into eukaryotes. And we think there's one optimization related to diffusion that really drives that uniformity in this one scaling relationship. However, metabolic power scaling shifts between prokaryotes and eukaryotes, the growth rate scaling shifts, the eukaryotes pick up a bunch of scalings that you obviously don't have in bacteria, like how many mitochondria they have. Jordan Oki has a really wonderful paper showing how many mitochondria you get as a function of cell volume. And so there's, lots of things are changing and then some things are staying the same. And both of those angles help us understand if we're onto the right theory, if we're onto the right constraints.

0:52:02.5 CK: So for example, if we think a dormant constraint comes with a limit that should imply a major shift at one scale, and we look at prokaryotes on one side and eukaryotes on the other side, and they have exactly the same scaling, then we were wrong about the constraint. Luckily that hasn't happened to us but that's a nice way to say we could easily falsify our theory if we'd pick the wrong constraints because there's certain things we would predict that just wouldn't hold up. You would move across transitions and either see shifts that you didn't expect or not see shifts that you did expect. And we think that's a sort of secondary test of these theories.

0:52:40.9 SC: So when you say words like, just to be nice to the audience, metabolic power scaling shifts from prokaryotes to eukaryotes, what would that mean? What are some synonyms for that?

0:52:53.2 CK: Yeah, that's great. So what I mean by metabolic power is just the total energy available to a cell. So this is how, it would be like the total food that you and I ate. For cells, it's the same. If they use sunlight, it's the total sunlight they're able to capture. If they eat other things, it's how much of that that they're able to eat, per unit time. So that's the sort of total energy budget the cell is the metabolic power. And then scaling is just this idea that, if you look across orders of magnitude in one feature, you get a linear relationship in orders of magnitude of another feature. And those could have different exponents, which would be the slopes of those different curves in that log log order of magnitude, order of magnitude space.

0:53:43.7 CK: And so when we say there's a shift in a scaling relationship of metabolic power, what I'm really saying is that as you go across this evolutionary divide, you see a change in the exponent from one class to the other. Specifically in prokaryotes, the exponent is greater than one, meaning, if you double in cell volume, you more than double the total metabolic power. If you go up by a power of 10 in cell volume, you go up by more than a power of 10 in metabolic power. In unicellular eukaryotes the exponent is sublinear, meaning that if you go up by an order of magnitude in cell volume, you go up by less than an order of magnitude in metabolic power. So the power per unit volume is increasing in prokaryotes and decreasing in unicellular eukaryotes. And that's a really fundamental difference that drives all sorts of downstream things that we can observe, like the growth rates, like the requirements for the number of ribosomes in the cell, all sorts of different features.

0:54:43.3 SC: So in eukaryotes you still need more power when you're bigger, but you need less than your extra bigness.

0:54:50.7 CK: Exactly. Exactly. You still need more power, but in a sort of gained efficiency away.

0:54:56.5 SC: Is there a simple way to connect that to the existence of a nucleus in some mitochondria?

0:55:03.9 CK: Yeah. So, it looks like, the nucleus is a more complicated story 'cause I think the nucleus is mostly about regulation and separating out certain parts of the genome from each other and protecting the genome from reactive parts of the cell. But the mitochondria seemed to have just come along for the ride of that. So there, I think there's a big debate about causality. People would say, well, the number of mitochondria have the same scaling as the metabolic power. So are they driving the metabolic power? And I would say, we don't quite know yet. We don't know what's being optimized, 'cause it could be that the number of mitochondria you have are just, optimized according to some other constraint. And that constraint is what's driving metabolic power and mitochondria come along for the ride.

0:55:53.9 CK: Or it could be some fundamental constraint about the mitochondria that require the metabolic scaling to be what we observe. But yeah, in general that's, exactly the sort of connection we try to draw. We say we have a physical constraint, maybe it predicts something like total metabolism. And then once we know total metabolism, we can write down a cell model. We can write down a model of cell physiology and ask what other components should scale in what way according to this metabolism. And, in bacteria that gives us a whole host of predictions and the same in unicellular eukaryotes.

0:56:31.9 SC: I can't help but mention as a slight aside that there's a lot of really cool work being done on artificial life, like simulated life in the computer, origin of life, or evolution and things like that. And generally the word metabolism is not there at all. These are just running programs and they have all the food that they want. But clearly, if we're gonna find some more ultimate end game understanding the constraints of metabolism in the real world are kind of gonna be a big deal.

0:57:00.3 CK: Absolutely. Yeah. And I think that's something we're starting to run up against more and more recently. I mean, it's interesting that in the early days of computer science, a lot of the considerations were really about, well, how much resources will this need as we scale up?

0:57:19.7 SC: Yeah.

0:57:20.3 CK: So how much time, how many CPUs, how will this scale up in a resource constraint? And then we got to a point where we were mostly interested in, well, just artificial life in the computer. We had enough resources. So then it's all just about how do you get the right sort of evolutionary dynamic? How do you get open-ended evolution? How do you get things that build up higher layers of structure and function? And there's a really wonderful bit of artificial life that was done there. And now with the emergence of AI and LLMs, I think we're back to really wondering about scale and constraints. There's a huge debate there about what will happen if we make the neural nets 10 times as big, and how many data centers does that take? And so I think we're back to thinking very clearly about scale and about ultimately the physical resources that you need to support artificial intelligence or artificial life or really sophisticated computer programs.

0:58:19.4 SC: Good. That was just an aside. I don't wanna get sidetracked by that. Like every modern podcast ends up talking about AI at some point. So like, it's an attractor, it's a constraint.

0:58:29.8 CK: We did it. We checked the box.

0:58:30.8 SC: We did it. Yeah. But we have, looming in front of us the next big evolutionary major transition. From unicellular organisms to multicellular organisms. Let me just first ask, how clear is it what qualifies as a major evolutionary transition? Is it just that, when you see it, is there a theory of major evolutionary transitions or is it just like, well, we look at the data and we have to explain what we see?

0:58:58.1 CK: It's a great question. So recently they've been called, for some of those reasons that you're mentioning Sean, in terms of the problems that raises recently, these transitions have started to become major transitions in individuality rather than major evolutionary transitions.

0:59:14.8 SC: I haven't heard that one. That's good. I like it.

0:59:17.6 CK: Yeah. Now I think individuality is a also a complicated concept. And so then we're into talking about, well, what is an individual and how do you define that? And people like David Krakauer or SFI work on entire theories of individuality. And that's a frontier area in and of itself. So I think that reframing doesn't quite solve all the problems with, how do you know one, how do you spot one? What's the right metric? I think generally what people are thinking about is you have extra layers of architecture or hierarchy where you've packaged together more lower level things into some higher level thing. And where that higher level thing in an evolutionary sense operates as a whole. So the selection that matters to it, the evolution that matters to it is really on the whole. Now many of the transitions are really murky.

1:00:13.8 CK: Where we have, for multicellular organisms, for example, we have lots of organisms that happily live as a single cell, and then come together in multicellular assemblages, and even have their own physiology and set of responses and behaviors in that multicellular assemblage. And then when life shifts, they go back into the unicellular level. That even happens in bacteria, bacteria going in and out of these biofilms and making complicated choices about whether to form a biofilm and stay in a biofilm. And so multicellularity is a great example of a really murky idea. I think many people would point to a major evolutionary transition in individuality, as multicellular where it's not so easy to go back and forth between the unicellular and the multicellular phase and where the multicellular really is being selected, the genomes of the multicellular are really being selected as a whole.

1:01:06.8 SC: Okay.

1:01:08.2 CK: So in our case, all of the features that we do, everything we do in our life if we're gonna have children results in a single copy, half a copy for each human, being combined with another half a copy to form a new organism. So eventually everything goes through one copy of the genome to go through development and form an entire other multicellular organism. So it doesn't matter that... I mean, my finger only matters in the sense that it eventually interfaces with that one copy of the genome.

1:01:42.5 SC: Right.

1:01:42.6 CK: So if I lose a finger, maybe I die because of that. Hopefully I don't lose a finger. But that little bit of tissue is its entire sort of connection is to this one genome. And it's hard for my finger to decide, actually, I don't really wanna hang out with this guy anymore, I'm gonna go do my own single cell thing and try and make my own living in the environment. It's sort of, all the cells in my body are committed to each other in a way that's very hard for any of us to defect.

1:02:15.6 SC: Right. Good. Okay. That makes perfect sense as to why a theory of individuality would be extremely relevant here. That's good. So, okay. Multicellularity itself, one might have guessed if one were completely ignorant of these things, that maybe eukaryotes that have more complexity inside, were more willing or able to become multicellularity 'cause They sort of had more degrees of freedom. But my impression that it, actually in the data, either one is possible, it's just that the eukaryotes kind of worn out empirically.

1:02:54.6 CK: Yeah. And I think it's, there's a whole debate there about what the, the eukaryotes have that makes them special for forming multicellular assemblages. People argue that it's the regulatory capacities that those cells have built up to say, regulate all of these mitochondria. So mitochondria are great but a bad mitochondria is just a parasite living inside of you as a unicellular eukaryotes. So you have to regulate its metabolism. You have to make sure it doesn't eat your resources. This is why many people think the mitochondria has such a stripped down genome and actually requires the nucleus to make things for it that get exported and imported into the mitochondria to regulate it and so forth. And so there's a huge...

1:03:39.9 SC: They're impressed.

1:03:42.9 CK: Exactly. And so there's a huge amount of regulatory machinery built up in eukaryotes. And some people think that's what's really essential for forming multicellular assemblages, where you can then differentiate and regulate cells of different types. In that same sort of vein, there's unique developmental capacities and unicellular eukaryotes that people think is maybe important. And so there it may be, in this case, there is a wall at the upper end of eukaryotes, but in this case it may be that it's not as hard to jump over the wall because the things that you're forced to do as a eukaryote actually make this multicellular step really easy. And in fact, we see many origins, many unique and separate origins of multicellularity in the history of life. And so that tells us something. It tells us that it's potentially a much easier step than other things, maybe, because all the stuff you build up as a eukaryote is exactly what you need to make the next transition in hierarchy. That may have been less true for prokaryotes who eukaryotes.

1:04:52.3 SC: Well, we did have a podcast with Will Ratcliffe. I don't know if you know his work, but he's doing...

1:04:56.3 CK: I know Will very well. Yeah.

1:04:57.6 SC: Good.

1:04:57.7 CK: Yeah.

1:04:57.8 SC: I mean, he's doing the experiment where he's stressing these yeasts and they decide that becoming multicellular is a good life survival strategy.

1:05:07.8 CK: Exactly. And I think that's just a really wonderful example of they set up a certain sort of physical selection and that selection is sinking or not sinking.

1:05:22.1 SC: Yep.

1:05:23.6 CK: And then because of that, you get organisms, these snowflake yeast that build up more surface area to sink more slowly, and they're the ones that get selected on in this environment. But yeah, I think the other, that's sort of two stories. One is, physical constraints matter, and the other is it seems pretty easy for lots of single cell eukaryotes to form multicellular organisms and then to start doing. And I think, in lots of Will's recent work, they're showing pretty sophisticated cell differentiation that you're actually starting to get cells performing different types of functions throughout the colony. And that's really exciting.

1:06:00.0 SC: And that does seem like something that eukaryote would pull off more readily than a prokaryote.

1:06:07.6 CK: Yes. I think so.

1:06:08.7 SC: Good.

1:06:10.9 CK: Now it is tricky. Bacterial biofilms do have lots of really interesting responses within them. And there's been debates about all sorts of higher order functions that they might be performing that look like cell differentiation. So.

1:06:29.3 SC: Okay.

1:06:29.8 CK: I think there, yeah. So there is some trickiness there. But maybe staying in that aggregate or getting irreversibly committed to that aggregate maybe is the hard part. Maybe once you get big and complex and differentiated, there's no way to regulate it in prokaryotes. I think that would be a leading idea. But yeah, I think that's still a frontier in my opinion.

1:06:56.0 SC: No, that's good. I don't wanna be eukaryocentric. I wanna give the prokaryotes their props. It's too easy to be in the situation that we're in, but that the callback to Radcliffe's work in the multi, what is it called? MuLTEE, Multicellular evolutionary, long-term evolution experiment.

1:07:16.5 CK: Yes.

1:07:19.8 SC: Like we said they kind of put constraints. There's a new physical external thing. And that's very analogous to what you say was going on in the actual origin of multicellularity back in the Snowball Earth times.

1:07:34.4 CK: Yeah, exactly. And so yeah, in sort of running with this idea of how physical constraints might set up evolutionary transition we worked with some paleontologists who were very focused on the fact that one of the huge expansions of multicellularity seems to happen just after Snowball Earth. And so Snowball Earth was this phase of earth history, where the earth was mostly covered in snow and ice, including the oceans. So the glaciers come, even the ocean glaciers, sea ice, come all the way down to the equator. And then people debate about whether this is a hard snowball or a soft snowball with some amount of open equator and some amount of open patches of water around the planet.

1:08:23.6 CK: But the point is it's mostly covered in ice and snow. And as an aside, the way that you get out of that is volcanic activity, which keeps pumping CO2 into the atmosphere with no way to take that CO2 up into the ocean and into life. And so then you rewarm the planet through a greenhouse effect, and that pulls you out of Snowball Earth, otherwise with no volcanoes, you just get stuck there. Which is a sad end to an interesting planet up to that point. So we said, well, you get this huge explosion of multicellular organisms just after snowball. That doesn't seem coincidental. And so can we take our models of the biophysics for unicellular organisms, say eukaryotes and simple multicellular organisms, and ask what would happen to those as the world cools down as you go into one of these snowballs? And so we sort of invented a sort of toy model organism to represent the multicellular, which is this swimming spherical shell of cells.

1:09:32.3 CK: Now we see multicellular in the modern world that look quite a lot like this. And so we said, that seems like a good model. It's hollow on the inside. It's these single cell aggregates that could decide to go back to being single cells if they wanted to, but they also can swim around and collectively feed and that sort of thing. And so then we said... And then you also have large single cells. And so now what happens as the world freezes? Well, temperature starts to go down first. So you first get this cooling ocean temperatures, and then eventually you start to form more and more ice. And when that happens, sunlight and what we call primary productivity, which is just how much sunlight is being captured and turned into biomass and producing sugars and fixed carbon in the ocean that then starts to go down as the ice covers the ocean. So in this cooling phase where only temperature is changing two things are sort of happening. The metabolic rate on that metabolic power that we were talking about is getting slower and slower because it depends on temperature. And as temperature cools that rate goes down.

1:10:42.4 SC: Okay.

1:10:43.2 CK: And what's interesting is that pushes single cell organisms to get bigger, and it pushes these spherical multicellular organisms to get smaller.

1:10:53.8 SC: Wait a minute, you gotta explain that.

1:10:55.6 CK: Yeah. So what's happening is the metabolic rate is decreasing. And so it allows these single cell organisms to live the same lifestyle as sort of a larger size because they sort of don't have as much metabolic demand 'cause Everything is slower. And then these spherical organisms as temperature decreases, they're facing increased viscosity. And so their life is a little harder for hunting. And so that's sort of pushing them to be smaller. But then as the nutrients plummet, as the sort of total biomass starts going down, this pushes the multicellular organisms to get bigger and the single cell organisms to get smaller. And that's mostly a hunting effect of sort of gathering resources effect for the large spherical things. And so we think the snowball pushes your multi cell organisms to get much bigger. And as you start to get much bigger you start to run into other physical constraints that would induce sort of more complicated geometries. And as you start to develop these complicated geometries you start to discover things like sponges, we think. And then you find this multicellular architecture that's really stable and really beneficial. And when the world thaws out, you keep it, you keep these multicellular around.

1:12:20.7 SC: So it's all because life got hard.

1:12:24.8 CK: Exactly, yes. Absolutely. Yeah.

1:12:27.3 SC: Life got hard and multicellularity was the solution, but of course, we had to get to the point where we were pushing the boundaries of what the single cell things could do before that became... We had to develop our single-cell technology well enough so that we were able to take advantage of a new modality in the multicellular world.

1:12:46.3 CK: Exactly. Yeah, exactly. Yeah.

1:12:48.8 SC: And then, okay I presume I can just sort of imagine that all sorts of new options open up once you become multicellular. I guess the differentiation is key. Different cells serving different purposes. On the one hand, you can't have that if you're unicellular. On the other hand, you don't need to have that if you're multicellular. So like that had to be a new innovation also.

1:13:13.4 CK: Exactly. I think that's a new innovation. And you can imagine in a simple way, what the calculus is of whether that's beneficial. So imagine I'm a sponge and I say, well, I have some reproductive cells, and they sort of do the reproductive bit of me, and I have to grow, and then I need things to feed that growth. So now imagine I am building a bigger, bigger internal sponge volume, and I just asked the question if I add a little differentiated cell that has Cilia on the outside that drives fluid flow through the inside of me through these filters where I can capture prey and then eventually digest them, do I get a return on investment for that? Every time I add one of these little cells with these little Cilia, which are just hairs that move water, is it beneficial to my total metabolism? And so that's a case where, maybe I have a few of those, maybe we're all that cell type, and then I say, okay, I make one of us reproductive and I keep all the other as sort of these gathering cells. Now the reproductive cell doesn't have to keep around all these cilia, which is not using, that's expensive.

1:14:36.1 CK: So I've differentiated through this one feature, and now all of these ciliated cells don't need to try and reproduce. So I don't need to think about the downstream consequences of what molecules they have to carry around to be able to reproduce. I get efficiencies out of both of that. And that's sort of what differentiation is doing. And I think that basic idea that, when you take one thing that does everything and you split it up into sub tasks that each do part of the thing but don't have to pay the cost of doing the other thing, but still get the benefit from it. I think that's an idea that runs through economics, through biology. It really is just sort of how economies of scale happen is through specialization and differentiation. And so I think that idea is as true for a city as it is for a little sponge. And it's just about when you specialize, you don't do all the things that you're not specializing in, so you don't pay the cost of trying to. And so, I'm not currently paying the cost of trying to become a really good guitar player, which I know is not accessible to me. And so I happily consume really wonderful music from other people and I get the benefit of that without needing to try to do it myself.

1:15:56.1 SC: That absolutely makes sense. But there is an extra layer of complexity here because in a car, the tires and the roof are doing different jobs, but they're also just made of completely different things. In the multicellular organism, all the cells start off kind of from the same basic material. So it's not a top down plan with different kinds of stuff. It's that the different cells in a multicellular organism learn how to respond to their environments and therefore become something different, which is kind of amazing to me.

1:16:29.1 CK: Yeah, I agree. I think development is astonishing. That you start off with one cell type and it divides and changes internal composition in terms of what proteins it has and what genes are turned on and you go down this whole developmental cascade. There are ways to replicate where you don't do that. Where you say, I'm actually just going to cut off a whole portion of the body that has all the different cell types already differentiated. And then all of those will replicate and rearrange and form a new functional differentiated body. There are a small number of organisms in different categories. Small number of large multicellular organisms in different categories are able to do that. Plants to some extent can, sort of amazing, you can grow the whole plant back from enough of the right tissue excised from an adult plant. And Planaria have all these strange things about how much they can replicate from bits. But certainly...

1:17:34.9 SC: We had Michael Levin on the podcast talking about that stuff. Yeah.

1:17:36.6 CK: Okay. Great.

1:17:37.4 SC: Yeah.

1:17:37.7 CK: Yeah. Perfect. Yeah. But there certainly are cases where you can do that and it... But it's interesting that it's not the norm that for certain classes of complex multicellular organisms. Mammals are a relatively narrow taxonomic group, but we don't have any mammals that are just fully cutting themselves in half, and then every cell replicates and builds a second portion of its same tissue, and you get a whole new mammal that way.

1:18:06.1 SC: Right. Yeah. Okay. But I do...

1:18:07.0 CK: No, plants are a large part of the world and there're Planaria area everywhere, so it's a little bit of how you count it, but yeah.

1:18:14.3 SC: So I do presume that the kinds of logic that we talked about at the very beginning of the podcast, where there are constraints that give you smallest things and largest things. I presume these can be applied to mammals just as well as to bacteria?

1:18:29.5 CK: Yes, absolutely. And I think that's actually the classic work that sort of re-kickstarted... Re-kickstarted, that's a double... It's a double rebooting, this sort of whole field and way of thinking. So those optimizations for the fractal geometry vascular systems in mammals and plants, got us thinking again about just how far we could take these physical constraints and the optimizations of those to understand what organisms are doing. And then we moved into the single cells in lots of different ways.

1:19:09.3 SC: Yeah, okay.

1:19:10.2 CK: But, yeah. Mammals have... It's a different set of physics. It's the physics of fluid flow through vessels. It's the physics of mechanical buckling under gravity and it's a very different sort of thing, but it still leads to predictions for how their metabolism should change with their overall body mass and what the structure of these vascular systems should look like, and how fast they should reproduce and how long they should live. And so there is a pretty well worked out sort of systematic theory there as well. It's just slightly different physics.

1:19:45.3 SC: Good. Very good. I think that that inspires me to wind up the podcast with what is probably a completely unfair question.

1:19:53.0 CK: Perfect.

1:19:54.6 SC: If the harsh conditions, well, let's not say harsh, changing external conditions on Snowball Earth prodded little unicellular eukaryotes to come together and start being multicellular. It's arguably, one could say that changing technological conditions on earth right now are creating very different conditions for life to exist and thrive and so forth. So I would guess that this opens up the possibility for dramatically, well, a new kind of a major transition in individuality.

1:20:37.0 CK: I absolutely agree. Yes. I have a whole bunch to say there.

1:20:41.3 SC: Please.

1:20:42.7 CK: First is, there have been many cases in the past where radical changes to the environment have led to these huge mass extinctions. And then following those mass extinctions, you get a, what's called a re-radiation or radiation of different organisms that then discover these new innovations and we're one such thing. As mammals that were small and in low abundance in the time of dinosaurs. And certain physiology allowed us to make it through this mass extinction and then to expand and become the organisms we are. So that's one side. I think that's totally a separate question from the ethics and considerations of mass extinctions from a responsibility perspective for humans. And it's also not guaranteed.

1:21:30.9 CK: So I think lots of mass extinctions don't get followed by a recovery. It's not clear that we always would've come out of Snowball Earth. It's possible that you find really interesting planets for a long time that then get into this snowball phase. The dynamicist, the planetary dynamicist will tell you that that alternative steady state having two steady states, one of which is a snowball is common for planets that likely can harbor life. And so I think all the time you could get into a snowball and never recover. So major transitions to planets, major shifts in planets may cause extinctions and be nonrecoverable. So they're definitely, that's definitely another possibility that doesn't always get mentioned with the extinctions lead to innovation and these new radiations and lots of cool new stuff.

1:22:24.6 CK: Extinctions sometimes just lead to the end of everything and that's really important to note. And then I think we're already in the midst of just an amazing time for, and I mean amazing with neither good nor bad, but really in the sense of amazement, time for evolutionary transitions. So many of us would say that cities represent a new type of organism, a new type of major evolutionary transition. It's certainly one of the major evolutionary transitions originally outlined in the first book on the subject along with human social groups. We've been able to look at the growth of companies using the same sort of models that we use for organisms. And so we're pretty comfortable talking about companies as entities as a new sort of entity that's on the planet.

1:23:18.1 CK: And I think different types of AI also represent a major evolutionary transition, if not the internet itself. In all of these cases, we're taking lower level components and combining them into new architectures. So cities combine humans in new ways, inside new architectures. The internet combines all sorts of things in new ways, in new architectures, as do LLMs, as do human organizations. And so I think we're in the midst of many new types of individuals showing up on the planet, which is really interesting. And then finally, I would say that also just how humans are changing the selective dynamic of the planet is also apparent. Ron Milo has some really nice work over a few papers showing how much of the planet's biomass is associated with humans, either our own biomass, or our crops, or the animals that humans tend to eat.

1:24:19.2 CK: And it's an amazing fraction, and it doesn't look at all like any other ecology we've seen in the past, over the course of the history of our earth. And so I think that's already a case where the human technology, our ability to organize our environment and plan ahead and construct our own niche has led to this really radically different ecology, really radically different selective forces. Who looking at the cow and chicken would predict these two organisms are gonna grow to completely surprising biomass densities 10,000 years from now. No normal ecology would predict that. And why cows and not other mammals of the same size is a totally contingent story that is based on human preference and human technology.

1:25:07.5 CK: So I think we are in this really interesting place where we've created totally new selective pressures for our ecologies and where we're creating totally new types of entities. And where that all goes, I don't think we have theory for, I think it's we're all figuring out as we go sort of thing. And there's lots of reasons to be concerned about all sorts of different dimensions of the future. It's a very interesting time in the long evolution of this planet.

1:25:39.1 SC: Well, we would... That's why I was thinking this an unfair question, but we would like a theory, we would like to be able to predict.

1:25:49.1 CK: Absolutely.

1:25:50.0 SC: If we do change the conditions in some ways, like could you, in retrospect, maybe someone could have been able to predict that some small set of livestock animals would become a huge fraction of the biomass given that human beings were technologically dominant?

1:26:09.9 CK: Yeah. I think maybe from a efficiency perspective, just to say, if you start to really understand one organism well, then you select on it and you pick it. I think there has, in general, and people have written about this, there is a trend towards humans eating less and less diverse things in time. And if you take a, say original natural environment and how many different types of things we ate, you move to agriculture, you move to industrialized agriculture, each of those steps, you're sort of eating less and less objects. So if you think about like in the United States, the fraction of the diet that is corn is large. So that's really, that's targeting one species, one crop, and why it's corn has historical and interesting reasons, but it's, maybe you could predict that you really do go towards monoculture for any technological species because of these economies of scale and efficiencies and so forth.

1:27:14.5 CK: At the same time, the other pressure that we see, and we have a whole project trying to understand this, is how and why human systems diversify in cities. But cities are doing the opposite thing, where as cities get bigger and bigger, you get a great expansion in the number of diversity of professions, types of restaurants, services provided, that sort of abstracted human landscape is becoming more and more diverse as the system size scales up. Maybe those effects feed back to the types of things we eat. So you can imagine, for example, and there have been movements like this in the US, but you can imagine people's preferences around diverse or heirloom or organic food starts to push things back in a diversity dimension in the same way that cities that have... The cities that are bigger and have more wealth, have a higher diversity of services. So you can imagine that feedback starting to push things in the other direction. And there I think it's complicated to say, well, will we end up with a mono crop or will we end up with an unbelievable rich amount of diversity, because that's what human preference is. I don't know. I don't have an answer to that. Yeah.

1:28:26.7 SC: Well, I guess it is one kind of interesting thing that people should keep in mind about constraints and how things evolve with respect to them. Even if there is some tendency to move in some direction and there's a constraint that prevents you from moving too far, it doesn't mean that everything runs right up to the limit. There is very often an equilibrium distribution that allows for some variety in there.

1:28:49.7 CK: Exactly. Yeah. And I think to your point about what we would predict, the surprise is that in every other ecology, we see one very particular distribution of body sizes. So we see a power law distribution of body sizes in lots of different taxonomic groups, whether that be mammals or bacteria or unicellular eukaryotes or plants. And we have good theory for why that's true. And so the question... So then the big surprise would be, okay, why do I have this one body size that's like a huge peak? Almost could be a signal that you have something, some strange organism around selecting on that.

1:29:32.5 CK: I think of leaf cutter ants as well that are growing this one sort of fungus that is likely in much higher abundance than you would expect from any other sort of ecological consideration of that. And those leaf cutter ants are also doing farming. They're also doing agriculture, happens to be underground. But it's the same sort of thing. They're gathering resources to feed one crop species that they really like. Whether you wanna call a fungus a crop or a livestock, is probably a debate for a taxonomist. But anyway, I think those sorts of surprises might almost tell you that there's some organism influenced environment through technology or intelligence or preference or something like that.

1:30:16.6 SC: And maybe one more reminder that we are not in equilibrium. We are in a state of flux, the modern world right now.

1:30:22.6 CK: Yes, absolutely.

1:30:23.4 SC: We'll see where we go. Chris Kempes, thanks very much for being on the Mindscape podcast. This was great.

1:30:28.0 CK: Thank you, Sean. Great being here.

[music]