Daniel Giraud Elliot

Transcription

JANE PICKERING: Good evening, everyone, and welcome. I've decided this lecture has perfect timing for a number of reasons. One is everyone can get home before the storm starts. And two, of course, today is Darwin Day. So what a perfect day. He would have been 205. So, yeah. Maybe. But anyway, perfect day for the launch not only of our Spring semester series, but the launch of the Evolution Matters series for 2014. My name is Jane Pickering. I'm the executive director of the Harvard Museums of Science and Culture, a member of which is the Harvard Museum of Natural History, which is the sponsor for tonight's lecture where we have a very eminent evolutionary biologist, Professor Rich Lenski who will be introduced in a minute. And I'm really excited to hear everything he has to say. So I just say the boring things, or some of the boring things first. One is that if you're interested in our program for the rest of this semester, you can pick up a program guide at the table to my right. And we also have a special post card for the Evolution Matters series if you would want to pick up one of these as well. And you can, of course, always sign up for our email list. We only use it to send emails out about museum activities. And again, you can sign up for that to my right. The other thing you can do on that table is work out how you can become a member of the HMSC as well and get lots of wonderful benefits. So I encourage you to do that if you haven't already. I would very much like to thank Doctors Herman and Joan Suit who have disappeared but are in the audience somewhere. And I can't see them. They're right there. I'd like to thank you very much for your generous support of this lecture series. And one of the wonderful things given their support is that we are able to videotape every lecture. So they are available online within a few weeks, and also all the lectures from previous Evolution Matters series are also available on the Harvard Museum of Natural History's website. So I encourage you to look at that. And also, just to mention, we have another lecture here next week-- next Thursday. A lecture sponsored by the Peabody Museum by Jason Ur on Lost Cities and Landscapes in the Heart of an Assyrian Empire. So I am actually going to introduce the introducer. And they introducer is Professor Jonathan Losos who is the Monique and Philip Lehner Professor for the study of Latin America here at Harvard. And he's also a professor of organismic and evolutionary biology, and a curator at the Museum of Comparative Zoology here at Harvard. He is the head of the Losos Lab, where his research team studies the behavioral and evolutionary ecology of lizards. And his most recent publication, "Lizards in an Evolutionary Tree," won him the Daniel Giraud Elliot Medal from the National Academy of Sciences. He received the Edward Osborne Wilson Naturalist Award from the American Society of Naturalists in 2009. And recently was elected a Fellow of the American Academy of Arts and Sciences. As with most Harvard professors, I could keep going in this vain for a long time, but I do would like to bring Professor Losos onto the stage to introduce our speaker. Thank you very much. JONATHAN LOSOS: Well, it's very fitting that Rich Lenski should be speaking to us today on Charles Darwin's 205th birthday. For those of you who have been attending the Evolution Matters series, you know that evolutionary biology is a very dynamic field, with lots of great researchers doing very exciting work-- many of them here at Harvard, but many elsewhere as well. Nonetheless, in the Olympic spirit, if we were to pick an all-star team of evolutionary biologists, there's no doubt not only would Rich Lenski be on that team, but he would probably be the captain. In recent years, few people have done as much as Rich's research to help enlighten our understanding about how evolution proceeds. His 25-year-long experimental study of microbial evolution that he's going to talk about today is simply extraordinary. And I could go on and on about this work and all the accolades it's received and all the publications in major journals, the thousands of citations in the scientific literature, all of the popular articles written about it-- one most recently in the New York Times. But I think to me the clearest indication of the importance of this research program is the unceasing attempts by the creationist to try to discredit Rich and his research program. Needless to say, they have not succeeded. And if you want an entertaining read, I suggest you Google Rich about this, and you'll find how he has turned the tables on them quite effectively. Now given the length and breadth of this research program, you might think it's the only thing that Rich does. But in fact, that couldn't be further from the truth. Rich has had a very broad and varied research career. Initially, he started out as a graduate student at the University of North Carolina working on, of all things, beetles and how beetles interact with each other in natural communities. It wasn't until we moved onto a postdoc that he started doing laboratory work, which he's been doing ever since. Not all of this work, however, has been in the wet lab, because in addition to the microbial work, Rich has also had a long term research program using very complicated computer programs to simulate evolution to learn about the genetics and about evolutionary processes, what are sometimes called the digital life. And this work has actually been extremely influential in helping enlighten our understanding of the evolutionary process. In addition, just to round things out, Rich also has interests in both the history and the philosophy of science. And he's an ardent proponent and practitioner of enhancing public education with regards to evolution. Finally, just a few other things about Rich, Rich got his undergraduate degree at Oberlin College, then went on to do graduate work at the University of North Carolina, as I said. He then was a faculty member at the University of California at Irvine for a few years before moving to Michigan State University, where he is now the Hannah Professor of Microbial Ecology. Rich has received too many awards and honors to list them all, but among the more notable ones, he has been elected to both the American Academy of Arts and Sciences just down the road here, as well as the National Academy of Sciences. He has received fellowships from the Guggenheim Foundation and the MacArthur Foundation, and most recently he was elected president of the Society for the Study of Evolution. So it's a great honor to have Rich speaking to us today. And he will speak as, the slide says, on Time Travel and Experimental Evolution. RICHARD LENSKI: Thanks, Jane and Jonathan. It's a real pleasure to be here. And Jane mentioned the birthdays. And then one of the remarkable coincidences of history, not only is it Darwin's birthday, but in the same day in the same year it's Lincoln's birthday. So you know one of the greatest statesman, if not the greatest of that century, and the scientist had the same birthday. You might think I'm going to put George Washington up, but in a more selfish vain, I was going to mention that if I'd been smart, I would have timed it to start it on the 12th, but in 1988 in February was when I started this long term experiment that's going to be the focus of my talk. But since it is Darwin's birthday, I thought it was worth starting out by discussing a couple of things that Darwin got very right before turning attention to a couple of things that maybe he didn't quite get exactly right. So, of course, Darwin's theory and theories encompass so many ideas it's not possible to discuss them all today, but the two most important, I think, were the idea of evolution-- the descent with modification. It was a genealogy to life as represented in this drawing from his notebook in 1837. So over 20 years before he actually published The Origin, he was beginning, after the voyage of The Beagle and so on, to coalesce his ideas and representing things as a tree in a way of-- I love this. Above the drawing he wrote, "I think." You know, there's his hypothesis beginning to appear. And then not only did he recognize the descent with modification, but he also, of course, recognized the process of adaptation by natural selection. And other people before him had certainly had evolutionary ideas, but getting this mechanism-- which she did also with Alfred Russel Wallace, who was sort of a competitor and simultaneous publisher-- Darwin really just, the overwhelming extent of evidence he brought to bear. So including, for instance, the finches of the Galapagos Islands, where he saw these differences in beak size. Although actually he was a little bit confused about it while he was on the Voyage of the Beagle, but later on including, with the help of John Gould, the artist who took these pictures-- or not took the pictures, but painted the pictures-- based on Darwin's collection, helped him sort out where these things had come from on the different islands. But Darwin wasn't perfect. One of his theories that doesn't really survive other than to historians of science was Darwin knew he needed a mechanism of inheritance. There had to be a way that these differences between individuals which he was so focused on would be propagated to future generations. And in one of his later books he produced this theory that has this word gemmules, and it's about circulating particles in the blood, somehow influencing what gets on to the next generation. So you can give him a little sense of some epigenetics-- very hot topic now. But certainly his genetics theory was wrong. And it took Gregor Mendel and his pea plants to figure this out. And even then, that was lost until people rediscovered these results several decades later. Another thing which he was wrong about, which really brings me to this subject of the talk today, is he thought evolution was too slow to observe. And he says that very explicitly in The Origin in a couple of different places. So here's sort of a very explicit quote where he says, "We see nothing of these slow changes in progress until the hand of time has marked the long lapse of ages. And then so imperfect is our view into long past geological ages, that we only see that the forms of life are now different from what they formerly were." So in a sense, this is a challenge to us evolutionary biologists. And that's why we're interested in time travel. I'm an experimentalist, sort of watching things as they happen. But I wanted to emphasize also that really all evolutionary biologists are interested in time travel, whether it's people who look directly back into the fossil record-- so this is a spectacular find of Neil Shubin in Chicago on this sometimes at Tiktaalik, but it's sometimes, I guess, called fishopod, this sort of transitional form. So paleontologists are certainly looking back into time and trying to understand the state of nature at different points in time. Or Jonathan might recognize these-- the beautiful anolis lizards-- the comparative method. Whether it's looking at DNA sequences to reconstruct the tree of life, whether it's looking at behavior or morphologies of organisms, comparing across all taxonomic scales from single celled two ourselves, the comparative method dominates evolutionary thinking. And while it looks at living organisms and compares them, by comparing things that live now, one looks back into the past and is able to formulate and cast hypotheses about previous states of life. So that's another approach where we're really engaged-- we the field of evolutionary biologists-- are engaged in time travel. And then there are those, like Peter and Rosemary Grant at Princeton pictured here, who watch the process of evolution as it occurs. They've spent decades-- good parts of each year on this very forbidding island in the Galapagos of Daphne Island-- where they have done intensive studies of, in fact, these finches that Darwin observed-- the Galapagos or Darwin's Finches, as they're now called-- documenting subtleties of variation within species, leading to evolutionary changes on the time scale of a few decades, similar, actually longer than the experiment that I'm going to describe, but a many, many decade-intensive studies. And because evolution is an observable process, in principle that means you can also do experiments with it. And that's, as Jonathan's introduction made clear, really what I'm going to be talking about. So the experiment I'm going to talk about in this LTEE-- that just what I always call it-- The Long Term Evolution Experiment. There's never room to fit it on. The fundamental question that's always interested me is how reproducible is evolution? How repeatable is it? And to me, this is really interesting because we know that evolution relies on random genetic events, giving rise to mutation or in some organisms, random recombination of genes during sexual reproduction, and yet we also know that natural selection-- the force that Darwin and Wallace first discussed-- that if you have the same genetic variation in the same environment, in principle, organisms could evolve extremely similar adaptations to one another. And so this tension between the idiosyncratic random effects and the sort of predictable, deterministic systematic affects has always just fascinated me as being at the core of evolutionary biology. And we know that this tension is also-- I mean there are other things going on. I'm sort of oversimplifying a bit. But on the one hand, this tension can at least contribute to the spectacular tree of life. This is Haeckel's picture from not long after Darwin's Origin. And then we also find these incredible striking examples of convergent evolution. So cephalopod mollusks and vertebrates have eyes that work very, very similarly, have very similar structures, and yet they're common ancestor didn't have that. This has evolved independently twice-- these extremely sophisticated adaptations showing, in fact, that evolution can be quite reproducible, given the clear selective advantage, at least in certain contexts, of organisms that have eyes that can gather extremely accurate information about the environments in which they live. So how do we explore this tension between chance and necessity? And I'd say there are two things we'd really like in a study system? Is one, if you're going to ask how repeatable something is, you would like replicates. You would like the ability to watch the same process multiple times. And you would like time travel. And in particular here, referring to that out of Jurassic Park. But the idea is going back to some point in the past in an experiment and being able to restart things to query the predictability of certain particularly interesting events. So this experiment that I'm going to spend the rest of the time on, there are 12 populations. They all started from the same strain of E. coli. And by the way, it's a non-pathogenic strain, so you don't have to worry about shaking my hands or anything like that. They're all identical except six of the lines turn out to make colonies of one color and six of another. And that colony difference allows us to compete the strains against one another and say to compete a red evolved line against the white ancestor. So I'll show some of that kind of data a little bit later. And what we do is somebody in my lab every single day, 365 days a year for now almost 26 years, somebody takes 1/10 of a milliliter out of a population and transfers it into 10 mils of fresh medium. And in that medium, the limiting resource for the bacteria is glucose, which is their sole source of carbon in energy. Every day when we dilute it 100 fold, E. coli's a fast grower, as you know, and so the E. coli can go through enough doublings-- which you can work out the math-- about six or seven doublings a day. They will rise back up to the population density where they've run out of glucose. And we just repeat this every 24 hours, 365 days a year. Now it'll become important later in my talk that, although I chose-- I was trying to make sort of the world's simplest evolution experiment here to get at this question of reproducibility. But for various reasons, in this particular recipe I used there is another carbon source. It's called citrate. Interestingly though, E. coli, as a species, going back to its original description cannot grow on citrate in the presence of oxygen. But I think you'll see that I'm going to go there eventually. So the experiment started in February of '88. So we're up actually today, in a small coincidence, it hit one of our [INAUDIBLE] birth. It's now at 59,500 generations as of today. Somebody in my lab reminded me of that. And if you calculate how big the populations, each one of these flasks holds many millions of cells. And we know something about the mutation rate of E. coli. Every one of these flask populations over the course of its history-- over these 26 years-- has literally sampled billions of mutations. So they all start out identical, but they're randomly generating all of these different mutations. And a very important feature-- when I switch from insects to bacteria, as Jonathan said, the ability to, you know, have so many generations. That was what was seductive to me. But really what's important about this system is we can freeze the E. coli, and the E. coli remain viable in the freezer and we could bring them back to life. And you'll see that is why that's so important in this system. And so this is what I mean by time travel. We're able to look at time going forward in the sense of these rapid generations, allowing me to stand here and talk about an experiment that's gone on for over 50,000 generations. We can freeze and bring back viable organisms from earlier generations. So you can see the bacteria may sit in the freezer and be able to do this time travel. Unfortunately, I've only been going in one direction, as you can tell from the color of my beard. I would like to be able to go back, but c'est la vie. So we can go back in time as well as forward. And when we go back, we can also go forward again. And I'm not explain. Later on you'll see what this slide is about. But when something really interesting happens, you might want to be able to distinguish between was it just something really rare that could've happened at any point in time, or was there a series of antecedent events than if you replay the movie just up to a certain point or whatever, you would get a similar outcome? So what kind of data can we see in an experiment like this? So one kind of data that really was the original kind and still serves us in so many ways, is what I call fitness trajectories. And the idea of a fitness trajectory is that we are measuring-- so many of you have probably heard there's this sort of nebulous concept in evolutionary biology that, you know, survival of the fittest or the morphed organisms leave more offspring, and so on. And those are all a good sort of Gestalt sense of it, but it's often very hard to really directly measure what a theoretical evolutionary biologist would call fitness. Well, what we can do is take our populations that have been frozen at these 500 generation intervals, bring them out of the freezer, acclimate them, get them sort of in the condition of the long term experiment, and then we can compete organisms that lived at different generations. And so in this graph here, I'm showing over 50,000 generations, the fitness measured relative to the ancestor of evolved samples from these different time points in that population. So after 50,000 generations, you can see, in this particular population-- I don't have the error bars and so on up there- but in this particular population, these bacteria are-- it's a ratio of their relative growth rates as they compete head to head. So after 50,000 generations in this constant environment, this population, the bacteria are growing about 80% faster than did their ancestor. So it's just, if nothing else that you take from this, it's a beautiful demonstration of adaptation by natural selection in something that is happening in real time that one can observe and quantify. But remember my question-- that I'm only showing you one population here. Remember my question was well, how reproducible is this? So what if we now, instead of competing-- so these are different time samples from one of the populations competed against the ancestor. I mentioned these red and white populations. What we compete one of the red populations against one of the white populations where they come from the same generation? In this case, this is what the data looked like. And here, the relative fitness is a little bit more arbitrary, because which one's in the numerator or denominator? But I just kept that constant. It doesn't really matter. Again, we start out with the relative fitness by definition and by measurement of one. For awhile, whichever strain was in the denominator seemed to have an advantage because the fitness of the numerator relative to the denominator went down. Something happened and this guy caught up and has maybe a higher fitness-- maybe some other things going out at the end-- but those of you with sharp eyes might know the scale is quite different here. What if we put these data on the same scale as the one comparing the fitness relative to the ancestor? And you can see that although they're not identical, they're following very, very similar trajectories. The fitnesses are almost indistinguishable, at least compared to the improvements that they've shown relative to the ancestor. So what else can we look at besides fitness? And I don't have time to show you all kinds. We've looked at cell size has been changing, a little bit counter-intuitively. All 12 populations make much larger individual bacteria than did the ancestor. There are various other things that have occurred in parallel. This one I just wanted to mention because it's kind of an interesting result. This is getting into some technical stuff, and maybe I tread lightly going too technical, but there's a new technique called transcriptomics. Those of you who are biologists certainly know about it, but it's new. But, you know, 10 or so years ago it became affordable for evolutionary biologists to look at something like this. And what you're able to do is look at the levels of expression of thousands of genes in a genome simultaneously. And what I'm showing in this panel over here is simply the red form of the ancestor against the white form of the ancestor for 4,000 genes. Some genes are very lowly expressed. Some are very highly expressed. Turned on at a low level. Turned on at a high level. This panel is showing one of the evolved lines after 20,000 generations against its ancestor. And it's still true that most genes that are lowly expressed are still low expressed. Most that are highly expressed, are highly expressed. But I think your eye will immediately tell you, and statistics will support it, hey, there's much more scatter here. Things are starting to change in terms of what the cell is turning on at high levels versus low levels. This is another one of the 12 lines. Again, much more scatter than relative to the ancestor. These two lines are 40,000 generations removed from one another. They're each 20,000 at the time we did this from the ancestor. But this is comparing the two independently evolved lines against one another. They are more similar after 40,000 generations of divergence from one another than either is to the ancestor. And that's because they've been changing their way in which they express their genomes in parallel fashion across the replicate-- across the many thousands of generations in these replicate populations. And by the way, this, as well as, I think, my next slide, shows one of the advantages of sticking with it for a 20-some year experiment, is all kinds of new technologies that nobody had even thought of are coming towards our experiment. So it's just been an absolute blast. I'm not super-- I'm not a techno-file, but the grad students and postdocs in my lab have brought these wonderful new methods. And one of those other methods that you no doubt hear about-- and, you know, people talk about the $1,000 or $100 genome for human sequences-- well, the first bacterial genome was not sequenced until 1995 and at the cost of many millions of dollars. And now it's possible for, I mean, we're really basically drowning in these kinds of data. But a few years ago, Jeff Barrick was the first one in my lab to sort of say let's take the genome sequencing technologies and apply it to a series of individual bacteria-- what we would call clones-- isolated from one of these populations. And again, I'm not going to go into all the detail, but the basic idea is we have the sequence of our ancestral strain, and all these different colors and symbols are showing, over the course of 20,000 generations, the accumulation of mutations in one of these populations as we go over time. So after 20,000 generations in this particular population, there were a total of 45 mutations in a clone isolated from 20,000 generations. These include everything from single base changes-- one letter in the DNA sequence-- all the way up to things that have inverted large portions of the chromosome. Now we know some of these mutations must be beneficial, because we see these heritable improvements in the fitness. What we would like to know is whether most of these that have risen to high frequency are beneficial. I mean, it's conventional wisdom, and I think correct wisdom, that the vast majority of mutations in most organisms are either going to be neutral, have little or no effect on performance, or they're actually going to be deleterious. But the ones that would survive and rise to high frequency, you would expect to include disproportionate numbers of beneficial mutations. So I mentioned that the E. coli genome has something like 4,000 genes. From the sequence data, I just mentioned only 40-some of them actually have a mutation in this particular lineage by generation 20,000. Yet what we found is when we look at these same genes across our other replica populations, again we see this signal of parallelism. Not identical mutations, but this is showing two genes where 12 out of 12 populations have acquired mutations in the same genes independently. It's not the exact same mutation, but if only 1% of your genes have mutations, it's extraordinarily unlikely that we would see this parallelism just by chance alone-- that the same genes would be changing over and over unless it is natural selection that is pointing to these genes and saying, these are genes that underlie the adaptation to this flask world in which we've embedded the bacteria. And because, so as I said this parallelism implies natural selection, and because bacteria are a model system for genetics, we can test that by genetically engineering strains that have just one at a time of these differences, it's a lot of work but when we do that, the vast majority of these pairwise strains that we produce-- the ones that rise to high frequency-- indeed individually confer significant selective benefits to the organism. I want to make an aside here. I don't know whether is Tammy Lieberman here in the audience? Hey, Tammy. So Tammy Lieberman is a grad student, or did you just defend? She just-- so she's a new PH.D, Dr. Tammy Lieberman. So she works here at Harvard, along with Roy Kishony in Systems Biology. He's out of town so he couldn't be here. But this is just a little-- I'm just going to interrupt my talk with a little vignette about some really important and interesting work that Tammy lead that was published in 2011 in Nature Genetics. And it concerns a real life why does parallelism can be so important in terms of our understanding of evolution. So it's not that human individuals with this inherited disease called cystic fibrosis makes them very vulnerable. And one of the leading causes of morbidity and mortality in these individuals is opportunistic pathogens-- bacteria from the environment getting into the lungs of these individuals and causing severe effects. There was an outbreak, unfortunately, among several dozen CF patients in the Boston area that started in the 1990s. And fortunately, the clinicians were very smart in saving these samples that spread through this group of patients being treated in the Boston area. And now along comes this possibility of looking at the whole genome sequences of these bacteria. So in this paper that was published in 2011, Tammy and Roy and their many outstanding collaborators on this paper, sequenced over 100 of these different isolates from different patients at different points in this small, but very important, epidemic. And what I would love-- I teach using this example and go into much more detail than I have time here-- but using the same logic of what happened reproducibly in different individuals within this population that was affected, they identified 17 different genes where the bacteria were reproducibly across different patients, acquiring similar, though not identical in most cases, mutations some cases the genes that were involved were rather, maybe I'd say obvious ones. The bacteria were often becoming resistant to antibiotics. Not surprising since they were being treated with antibiotics. But also, they uncovered many other genes that were showing this evidence of parallel evolution. And that is telling us what matters to the bacteria in terms of their ability to thrive in the lung environment. Now, that paper was only published a couple of years ago. But the idea of this is it may suggest new therapeutic strategies, new ways of minimizing these infections, treating them, or whatever. So it's just a beautiful example that has real world implications of taking this kind of evidence of parallel evolution. And again, this is, by the way, evolution in action. Some of it, again, was reconstructed after the fact. But the dates at which these strains were isolated are well-known, and they could create a phylogenetic tree and so on. So it's absolutely beautiful work from Dr. Lieberman and colleagues. Now back to my experiment. This is-- I've shown you that we see-- I presented it in the exact opposite order. But we see these parallel mutations-- similar but not identical mutations accumulating in the populations. They're giving rise to similar but not identical changes in terms of the way in which the cell is expressing the networks of the thousands of genes there. And that's giving rise to similar but not identical trajectories in terms of their mean fitness. So it's as though-- so actually, I should explain this metaphor. Evolutionary biologists often talk about what are called fitness landscapes or fitness peaks. And the idea is that natural selection is driving populations to climb mountains. Dawkins called it Mount Improbable. And the idea is that organisms are, through mutations, getting closer and closer to some optimum within a given environment. And it's well-known that when environments change, this all gets much more complicated. I should also mention it's the same for any of you who have a physics background. It's like an energy minimization physics, except evolutionary biologists are more optimistic, so we're thinking about maximizing fitness rather than minimizing energy. But it's the same kind of mathematics there. So it's as though, to this point, I said we take an arbitrary ancestral strain of E. coli, we plop it into this laboratory environment that's obviously not its natural environment, and the populations are climbing up Mount Glucose. They're getting better and better at this very specialized environment in which we've given them. They're not doing it exactly the same, but they're doing it similarly. Well, there have been a number of very interesting exceptions to that. But what I want to do now for the rest of the talk is focus on a particularly striking exception. And after 30,000 generations-- so about halfway to this point of the experiment-- one of these populations suddenly said, hm, there's something else to eat besides that glucose. You know, they had always gone to bed after they ate their glucose. You know, didn't know there was desert. Nice citric, lemon-flavored dessert sitting there. But one of these populations evolutionarily discovered it. I was quite sure we had gotten a contaminant into the medium, so I was all worried about that. But in fact, we sequenced-- well, first we did a number of tests, and then subsequently we've sequenced it. And very clear, this is a new lineage within this population that has overcome this key phenotype-- this key trait-- that had defined E. coli as a species going back to Escherichia's earliest work to naming who gave rise to the genus. Now, there are a couple of ways about thinking about why one population-- and by the way, it's 60,000 or 59.000 and some generations. None of the other populations have figured this out still. So there are a couple of ways of thinking about what might be going on. One is remember I said there have been billions of mutations. That turns out the more mutations than there are sites in the genome. But not all mutations happen at the same rate. And in particular, maybe this required some extremely rare genetic mutation where you had to like invert some segment, and they had to do it at exactly these particular break points. Alternatively, it might be some perfectly ordinary mutation, but that only creates the new-- only generates the new citrate phenotype after some of the other mutations that are similar between the populations in many cases, but not always a not identical, after some of these other mutations have created what, as evolutionary biologists would call a genetic background in which now, ah, you're one step away. So it's as though they might be climbing towards-- they're all climbing glucose. But in the course of climbing Mount Glucose, perhaps one of the populations, just like a mountain climber might discover a new higher peak, depending on which way they go in their ascent up, you know, a k2 or whatever. And that would be sort of what we would call historical contingency or path dependence-- the likelihood of getting to this novel solution depends very much, not just on one event, but on a series of events. And the late great Harvard paleontologist, philosopher, historian of science, in his book Wonderful Life-- well worth reading if you haven't read it. Some people wonder if my whole experiment took inspiration from this book, but '89. Experiment started in '88. But it's still a beautiful quote. And the whole book is about the Burgess Shale and how reproducible evolution would be, and what if we could go back in time. So this was the quote. I should just read it instead of trying to explain it. He says, "I call this experiment 'replaying life's tape.'" And by the way, it already shows you evolution-- at least cultural evolution-- is happening really, really fast. When was the last time anybody in here had played a tape? So "replay life's tape.'" You press the rewind button and go back at any time and place in the past. Let the tape run again and see if the repetition looks at all like the original." And what he imagined, you know, if you had an interplanetary budget, you know, if you had replicate planet earths and you could go back to this time of what's called the Burgess Shale in Canada, where some of the earliest body plans of many of the animal phyla are present in the fossil record at about the earliest point in time, that's what he had in mind. Wouldn't it be really cool to be able to replay that? And he says we can't perform the experiment. That's the bad news. And of course, we can't on that scale. But conceptually, we can within this experiment I'd talked about. Because we have this frozen fossil record. And they're not just fossils. They're living organisms. And we can go back in time. So now I'm going to start explaining this slide that I showed you earlier. It's of Zachary Blount, who, when this work began, was a graduate student in my lab. He continues to do pretty much all the work I'm going to describe from here on out as a post-doc. And the idea was we wanted to be able to go back into the freezer and take, in this lineage that started from the ancestor-- you know, we have a sample from 500 generations, and 1,000 and 2,000 and 3,000 generations-- take individual clones and see whether he can re-evolve the citrate utilizing ability, starting from different time points. So in each case, he's starting with individual cells. So it's not like kind of was there some rare variant in the population. He's trying to ask about the potential to make this transition. And since we wanted to replicate, in a sense, the experiment I showed you-- and that was actually done in these glass flasks I showed you-- well, it turns out it was really, really hard to get this trait to re-evolve, and so he wanted to scale this up. And this is about 1/3 of the Petri dishes that Zack used over a course of a couple of years as one part of his dissertation. The good news is there were almost no colonies on most of these plates, as you'll see. But the idea was you can now grow up in a test tube and then centrifuge down and get like a billion E. coli from this clone at this particular point in time. And you put it on a Petri dish where there is nothing-- no carbon source, no glucose source, nothing but citrate-- and you let it incubate for a month, and you ask, is there a colony there indicating? And then you check it out to make sure it's not a contaminant, et cetera. You check it out to see can you re-evolve from different starting points along the lineage that eventually evolved this new phenotype. He ended up testing over 40 trillion cells outside of the pharmaceutical industry. I think it's the largest mutational screen. Nobody's challenging me when I say that in front of a bunch of microbiologists. Well, 0 of-- but I'm actually emphasizing here-- 0 out of 1- trillion of the ancestors when he replayed this experiment. And by the way, he couldn't handle all those Petri dishes of once. So those he was storing in the cold room. So he ran this experiment in a series of what statisticians would call blocks. And there were always the ancestral controls. And never, ever did he see one of them being able to evolve this new function. So I joke, you know, well, you have the tightest confidence interval around the value of 0 in the history of genetics. But well, did he get angry? But no. His experiment was a huge success-- a smashing success. So we have fun in the lab. And I saw some young people come in. It's really important that science sometimes comes across as so technical, but it's so much fun, especially when you're working with great people like Zack and all the people I've had at my lab. Well, what Zack found, when you crank through all these 40 trillion cells, never did the ancestor or any of the clones that he isolated before about 20,000 generations-- nothing magic about that dividing line. We don't have a lot of statistical power to say exactly when it is. But none of the early clones ever could make this transition. But in fact, he found 19 additional citrate-positive mutants. All of them came from after generation 20,000. And in fact, most of them came from a couple thousand generations before that new phenotype-- that new trait, the ability to grow on citrate in the presence of oxygen-- had evolved. So that supports, not that it's a rare event-- although we still think, given those low numbers it's a difficult transition to make-- but also that it's saying that, in fact, this historical path dependency-- this historical contingency that Gould argued would, in fact, be important in the history of life at this huge scale, also at this relatively modest 25-years scale, and in this case is playing a very, very important role-- that some things are improbable not just because the individual mutational events, but because they require a sequence of events to set the stage for the transition. Now, I mentioned earlier that we have the ability now to do whole genome sequences. And we went from being able to look at just a few genome sequences from one population to now, by the time this paper came out, we were looking at 30 genomes. Now back home, I've got to wrestle with several hundred from another project. But what we're able to do-- I mean, this is a really neat thing. It's a little bit like some data you might see for HIV and influenza virus where literally, you know, this is the ancestral sequence, and we have the number of generations marching up here. And what we have to show here-- each one of these little lollipops-- is a genome that we sequenced. And we were able to see, in fact, that within one of these three lineages-- this all in the same population. So it's pretty interesting that we're seeing divergence of lineages here. But within this population we can identify when this particular event gave rise to the first citrate-positive cell. And we can break this sort of evolutionary process down into what we call three stages-- potentiation. What were the mutations that set the stage for then being able to get one more mutation to become citrate-positive? That first mutation that makes them citrate-positive, we called that sort of the actualizing or actualization mutation. And then, of course, when any new trait evolves in an organism, it's not like it springs. You know, the first airplane with the Wright Brothers too. You know that didn't-- that's not a 747 or whatever. These things always there's room for improvement and refinement. So we, interestingly, still haven't figured out the full suite of mutations that are required for these early steps of potentiation. But what I can tell you is this is the specific mutation they made the first cell go from citrate minus to citrate positive in this population. And it's interesting. It's not a point mutation. It's a duplication in which one copy of a region of the genome ends up getting a second copy immediately adjacent to it. It's called a tandem duplication. And at first you might think, ah. Now with two copies there's a little bit more of some protein. It's actually the transporter that allows E. coli to bring citric acid into the cell under anaerobic-- no oxygen environment conditions. But in fact, what really matters here-- I'm not going to present the detailed genetic evidence-- is it's not the dosage of having two copies. It's that you've essentially rewired the genome. Because the expression of a gene of a function depends not only on a protein, but regulatory sequences that are nearby it. And when you make a duplication, what used to be downstream of the original copy, puts new DNA upstream of the second copy. And we were able to show that it's this rewiring of regulatory and structural genes, what some biologists call a promoter capture, which actually, in this case, generated this new phenotypic property-- the ability to grow. So the last topic I'd like to discuss is whether we now have a new species. And like so many things in science, and perhaps especially in evolution, a lot depends on definition. So I'm going to quickly look at several issues and then get to what I think is the coolest and most interesting of those. So one thing in terms of the way evolutionary biologists think about species and speciation is whether you have a situation in which one thing replaces another and the old thing is gone-- and that's called anogenesis-- or do you have a case where you actually have a bifurcation, where one new thing arises and the old thing-- the old trait, at least-- persists in the population. And I haven't presented to you the evidence, but maybe you might have noticed it in that phylogenetic trait. But it doesn't matter. It out this was very clearly a case of splitting of lineages. Even after the citrate-positive lineage evolved, the citrate-minus bacterias stick around for many, many thousands of generations. And to any ecologists in the audience, it's a stable coexistence. So you could start with them at different ratios and they converge to kind of a coexistence. So that's good. That fits kind of the way we'd like to think about species. It's also not just like the cells getting bigger or smaller. It's a novelty. It's, you know, sort of a discrete change. You didn't have it. Now you've got something. Well, that happens pretty often in evolution. You become resistant to an antibiotic or something like that. But this is a pretty meaningful novelty-- the fact that it affects a trait in one of the most well-studied organisms on Earth that, in fact, is not within what's usually considered part of its repertoire. But those of you who are evolutionary biologists would say well, what about this thing called-- what do evolutionary biologists really care about? And they care about this biological species concept, which really, in many ways, goes back to Darwin, although he didn't put a label on it. But it was formulated and made much more precise, again by a Harvard faculty member, the late Ernst Mayer. And he wrote, and he gave it this name of the Biological Species Concept. The exact wording isn't that important here, but I'll read it. "Species are groups of actually or potentially interbreeding natural populations, which are reproductively isolated from other such groups." Well, mine's obviously not a natural population, but the key issue here is reproductively isolated. And dogs don't mate with cats. But things that don't mate is only one way in which things can be reproductively isolated. They can be reproductively isolated even if they mate, but the intermediates are not well-suited to either of the environments of either parent, for example. But this is a problematic definition, even though it's the one evolutionary biologists love, for the most part. I'm sure Jonathan and others can note people who aren't wild about this. Bacteria are asexual. They reproduce by binary fission. Well, in nature, bacteria, even though they reproduce asexually, they exchange genes-- at least many of them do-- when viruses move genes around between different bacteria. And there are also things called plasmid-- extra chromosomal elements-- that allow the movement of genes. Well, we don't have any of that either in our experiment. That's one of the reasons they made the experiment as simple as possible was to the particular strain I work with, there are no viruses in the medium. There's no plasmids in the cell or whatever. But again, I'm interested in the speciation questions to probe our understanding of evolutionary systems, not because I really care about the names of species, at least in my case. What if we introduced horizontal gene transfer? What if we now sort of take our system and allow horizontal gene transfer to occur, would we see a barrier to gene flow? And in particular, the kind of barrier to gene flow I'm thinking of is this idea that maybe intermediates between the citrate-positive and citrate-minus type might be maladaptive-- they're neither fish nor fowl. And that if the intermediates, the hybrids, are maladaptive, that implies that those would be sort of absorbing states for the genes moving back and forth. And so that would be an incipient reproductive barrier. As I said, would produce a barrier to gene flow. So the idea is that if this is sort of our cartoon version of the world-- they're climbing up Mount Glucose, one of the populations has a mutation that starts to bring it into the domain of attraction where they can adapt to citrate. Well, when they first make this transition, there really isn't an intermediate. It's just the first mutation-- that actualizing mutation. But over time, presumably, they're becoming better and better at using the citrate. And so this valley between the glucose and the citrate might become deeper and deeper over time. Although we don't really know it. Maybe there's some ridge that connects them. That's why this is only a prediction. But the idea- the prediction is-- that as this new function gets better and better after it emerges in the subsequent thousands of generations, does the valley become deeper, and the barrier to this introduced gene flow would become stronger. So the first question is whether this new function does, indeed, get better over the subsequent thousands of generations after they became citrate positive. So I showed you early on some of these fitness diagrams. But in this case, I'm not competing-- but first of all, I didn't do the work. Zachary is not competing the later generations against the ancestor. What is competing is the later citrate-positive lineages against the very first citrate-positive lineage. And low and behold, you can see, in fact, that the bacteria, not only did they evolve this new function, but not surprisingly, once they get this new function, they're getting it refined, improved, over time. Well, we can do a genetic trick. We'd really like to be able to just mix the genomes. And George Church here at Harvard and others are working on techniques to do that sort of thing. But what we've done so far is what we can take is this rewired module of structural and regulatory genes, and we can put-- we can take genotypes from different time points after they evolve the citrate trait, and we can take away that key mutation-- that actualizing mutation-- and put the ancestral version back in. So we keep all of their refinements, except we pull the rug out from under them, because they no longer can grow on the citrate when we take away that first mutation. So they are, by definition, confined to this ancestral niche. And we've confirmed, of course, that they are citrate minus. And the question is, are they actually becoming less fit in the ancestral niche, as they have become more and more fit in this new citrate-using niche? And this is the same kind of experiment where we take clones after they've evolved the citrate utilization, and systematically replace that key mutation with that ancestral form of that, and they're not getting more fit. They're getting less and less fit over time. So this valley between them is getting deeper and deeper. So in fact, there is something that conforms to this very fundamental idea of a barrier to gene flow, that even though it's not part of our experiment, we can show that this phenomena is happening even in this very simple experimental system. So I would argue that yes, the citrate users are becoming a new species-- are a new species. I don't really worry too much about the words. But we clearly see this process that Mayer and others have emphasized as being very important to how we think about the origin of biological diversity, and how that origin goes from just a mutation to being something that gets locked into different evolving lineages. So in summary, I've shown you that experimental evolution allows us to travel in time and explore this tension between predictability and idiosyncratic outcomes. We've observed a lot of parallel evolution, though again, I want to emphasize it's not like parallel that the exact same things are happening. But there's still this big signal of many things in evolution at the level of this experiment, certainly appear to be quite reproducible, both in terms of the traits of organisms and their genotypes. I briefly mentioned the wonderful work by Tammy Lieberman an Roy Tishony, showing that the same kind of logic can inform our understanding really of many systems, but in particular, a very important pathogen. We also saw this new trait leading to the appearance of a new function in this very well-studied model system, and that that novelty-- this evolutionary novelty-- was not just sort of a one-off rare mutation, but in this kind of Gouldian way of historical contingency, it was the accumulation a kind of a meandering path-dependent evolution. And it shows signs that this is not just some freak of nature, but is, in fact, in some ways in a laboratory sense, becoming a new species. So I've told you about work that's extended over 26 years. I'm obviously not doing the transfers today in my lab. It's been awhile since they've let me in the lab. But over the years-- and this is only a partial list-- I've had so many wonderful grad students, post-Docs, undergrads, technicians. In fact, I'd like to give a special shout-out to the technicians, because I've been very fortunate over these 20-some years to have a series of three successive technicians who really help keep this experiment going, as do in many ways the people shown here. The National Science Foundation has been a funder of this. And this latest work on speciation has been funded through a grant to Zachary Blount by postdoc from the Templeton Foundation. And we also do have an NSF-funded center. And Jonathan, in his introduction kindly mentioned it, where it's not just me-- it's many. We're working with computer scientists and engineers who are trying to take the principles of biology and of which evolution by natural selection is one of the key, and apply them to robots and to computational systems and so on, so that there's a deep, not just using computer science as an aide to understanding evolution, but also taking ideas from evolution and other biological ideas and applying them into technological realms. So I will stop there. And if there's time, try to answer questions. So we do have time for questions. AUDIENCE: I was wondering. You said that this line developed the citrate, they totally used the citrate. But then when you tested it or put and cohabitated it with other lines that couldn't use the citrate, it's not that it dominated, and, you know, the other one, they sort of coexist. How do you understand that happening? Why didn't it-- why doesn't it just take over? RICHARD LENSKI: Yeah. So I think actually you may have set me up-- yeah. Stable coexistence. And I didn't even pay him. So this is the kind of evidence we do. What an evolutionary biologist wants to do in these conditions is mix them at different starting ratios. So this is across several orders of magnitude. And although most of the cells end up being citrate positive, there's about, when this first arose, there was about 1% to 10% of the populations remain citrate-minus. And their advantage is that they grow a little-- they can't use the citrate. So this is a growth curve of the citrate user. But the non-citrate user gets a faster jump. Every day remember, they go from having exhausted the resources in the previous day to getting transferred back to fresh medium. And the ones that have grown on glucose at the end of the previous day-- because they can't use citrate-- are able to jump back and use that glucose earlier the next day, and so they get a head start relative to the ones that have first used glucose, and when they run out of glucose, switch to citrate. Well, the next day, they have to switch back to glucose. And we don't know the full physiological details, but that's a sufficient trade off. So that's sort of the buzzword in evolutionary biology-- a negative correlation or a trade off in different performance measures-- that allows this stable coexistence. Yes, Arum. AUDIENCE: If you could travel back in time, knowing your current development and pick your 1988 help, what would you be able to tell them about the pliability that citrate [INAUDIBLE] would evolve. RICHARD LENSKI: Great question. And, you know, hm. It's very funny to go back and actually try to think what I was thinking at that time. And I was aware the citrate was in. When I was a postdoc I was working with the same recipe. And I tried to take the citrate out of the medium. And even though the E. coli is not using it as a carbon source, it chelates trace amounts of iron, which are necessary for E. coli to grow. So if you don't put the citrate in the medium, the bacteria are generally, from day to day, subtle fluctuations in the environment seem to matter much more. What I vaguely remember thinking was that, oh-- well, and I tried taking the citrate out of the medium again earlier before I started this experiment. And it wasn't a good recipe. So I think I thought one of two outcomes would be most likely. Either hey, it's actually not that hard to evolve. They're all going to figure it out. It'll be kind of a cool result. Or well, it's just not in their-- it's too far away from their toolkit and none of them are going to happen. Both of those are plausible outcomes. And I think, in some sense, I lucked out that it's a very interesting intermediate case where it's neither hopelessly impossible, but not commonplace. And that becomes, I think, fertile ground for the kinds of questions that we've then gone on to ask. So I guess I would try to convey that information to my younger self. Yes. In the back? AUDIENCE: So [INAUDIBLE] the rate that fitness and movement kind of a changed your mind, and some kind of a spike in that [INAUDIBLE]. RICHARD LENSKI: Yeah. So let me, and just in case people didn't get the question was how does the rate of improvement in these populations change over time, and do we see a spike, in particular when it involves the ability to use citrate? So we certainly do see a spike when they evolve that ability to use citrate. And in fact, it really becomes very, very hard to measure, because it just happens so quickly when that first cell-- maybe not the first one that became citrate positive, but then when they picked up another, the first of those refining mutations it kind of-- so it's sort of off the scale of all the other things I showed you. With respect to sort of the trajectory, what does it look like? We had a paper come out late last year in science. I didn't have a chance to talk about it today, but in my younger self, I fit these curve fits to the data of the fitness trajectories so that I used what's called a rectangular hyperbola, which has an asymptote. And that suggests that oh, they're finding these peak, and if you read the evolutionary biology, people are always saying, you know, they get to the fitness peak, as though that's something that can happen in some reasonable timescale. When a graduate student, Mike Wiser, went back and pulled all these strains out of the freezer and did very, very careful measurements, what we found was actually the data are very well fit by what's called a power law. A power law is one in which the rate of improvement gets slower and slower, but it doesn't have a mathematical asymptote. It can go on increasing forever. And then with a theoretician in my lab who's got a physics background, he was able to actually develop a model that's based on very much first principles and a couple of other things we know about our system and show that, in fact, that's what we should have been expecting all along. On the other hand, I guess I'm glad I didn't tell my early self, because 25 years later then we get a science paper out of it. So that's good. Yes. AUDIENCE: You talked about some other fitness parameters like the size of the cell and so forth. What are some of the other things your people notice? Like reproductive rate, perhaps? RICHARD LENSKI: Yeah. AUDIENCE: Things like that. What else did you look at? RICHARD LENSKI: Yeah. So I'll make a little tiny prelude before answering that. And yank me back to it if I somehow stop at the end of this, but especially in 2009, which was the 200th birthday of Darwin and the 150th anniversary of the publication of The Origin, I would often give talks after other evolutionary biologists like Peter and Rosemary Grant. And they would work on Darwin's finches. And that's an organism that really, in the scheme of things, is our close relative. They could empathize with what that organism has to do. It has to eat and it has to mate if it's going to leave descendants. And they were able to measure differences in beak morphology and plumage coloration and attractiveness to mates. But for them, the hard part was showing are these really subtle differences, do they have a genetic basis? They were able to establish that. And do they really matter? And they were able to show, especially in particularly stressful years where there hadn't been much rain, it really did affect their survival in reproduction. Hey, I come along. I work on these microbes. It's really easy to show that there have been these improvements. But what has changed? What can we measure? And so for a long time it was mostly these fitness curves and cell size. Also cell shape has changed. The cells are still rod-shaped like E. coli, but they tend to be a bit more-- their ratios have changed, so they're a bit more spherical. Most of the populations are less rod-shaped, I guess I should say. We did look at these kinds of questions of growth. So they're exponential, what would be maybe called fitness components. How do you break fitness down? All 12 lines, their maximum growth rate is much higher than was the ancestor. All 12 lines have also shortened this period-- the amount of time it takes to go from having starved at the end of the previous day-- to growing the next day. So they get a jump start. They've shortened that lag phase. We can also do experiments where we say, what if we measure fitness, not in the environment where they evolved, but what if we change the sugar on them? So not during the evolution experiment, but as part of the assay environment. And we find very interesting things there. So there's another sugar called maltose. Maltose is diglucose. Once it's down in central metabolism, it ought to be the same. You'd think that their fitness improvements would largely carry over if you just substitute maltose for diglucose-- which is maltose-- for glucose. And the answer is actually, suddenly these lines that show very similar fitnesses in the environment where they evolved, you reveal all this latent variation-- all this variation that hadn't been exposed to selection-- was free to wander. So some of the lines-- this was work we did relatively early, so I don't know whether it still holds up. But at least at some point in the experiment, some of them, while they were gaining substantial amounts of fitness on glucose, were actually losing fitness on maltose, suggesting that some of the very few differences-- and it turns that glucose transport and maltose transport use different pathways. And that's probably a big part of the difference there. So I hope that at least begins to sort of-- some of the traits we've looked at. I think, you know, going forward. I tell people this is the experiment that keeps on giving. It keeps on giving because rare events and a lot of evolutionary questions you do need a lot of time. It keeps on giving because the bacteria are doing very interesting things, and especially all the terrific students and postdocs and collaborators I've had. And I think I really want this experiment to continue long after I'm gone. So if anybody knows people who would like to endow an experiment into perpetuity, we can make some precise predictions. And it's not a costly experiment, so keeping the experiment going is fairly easy. The hard work will be analyzing things. And I think with systems biology, with metabolomics, and so on, I think kind of trying to integrate all these things we've been able to measure to date, and begin to understand what really matters, it's still a little bit opaque. The E. coli is still a bit, despite being such a model system, is a bit of a black box. And many of the mutations we've seen, when we do the whole genome sequencing, have been in genes that are a little bit-- they're complicated ones to understand because they are major dials-- major switches in E. coli metabolism. So something called DNA topology, which is how tightly wound the DNA is inside the cell. Hey, that's been evolving in 12 out of 12 of our populations. That's potentially affecting all kinds of things. There's something else called the stringent response, which is how E. coli deals with transitions from growth to starvation. That's been evolving in most, if not all, of the populations. So how that all works into sort of a model of something more like Darwin's finches of the phenotypes that matter, I think that's a big opportunity over the next decade or two or three. We had several over here. I'm not sure what-- AUDIENCE: So [INAUDIBLE] you have this threshold around 20,000 generations so that [INAUDIBLE]. RICHARD LENSKI: Yes. AUDIENCE: And [INAUDIBLE] evolving [INAUDIBLE] citrate, or do you think [INAUDIBLE] evolving that threshold to the point where they were going to do something [INAUDIBLE] in the [INAUDIBLE] generations around that [INAUDIBLE]. RICHARD LENSKI: Yeah. So let me repeat the question for those who might not have heard was since something was changing in this particular population that evolved the ability to use citrate, was it changing specifically so that with the right mutation it could use citrate, or was it in some sense doing a maybe more drastic reorganization so it might have been able to jump to x, y, or z that aren't citrate in the environment? And let me answer that a couple of ways. So first of all, I think the answer is that it was just the luck of the citrate. I don't think they were doing anything more general, but there's a 10-year-old who emailed me-- her mother emailed me yesterday. Is she in the audience? There she is. She asked an amazing question after reading the Wikipedia article about this experiment. One of the other cool things I did. And Carolyn or Caroline was the name. KAREN: Carolyn. RICHARD LENSKI: Carolyn asked this wonderful question, and her mother Karen. But some of them, I didn't have time. It's one of the things, in the interest of time I cut. But it turns out several of the populations in this experiment-- in fact, at 50,000 generations, six of the 12, have evolved changes in the actual rate at which they mutate. They have evolved changes in DNA repair and DNA metabolic pathways that change the rate at which evolution mutations are arising. And so Carolyn asked this wonderful question of whether, in fact, that this was one of those populations. So that actually, it had become more evolvable because the mutation rate had gone up. And the answer is no. This population, when it evolved the ability to use citrate, was still at the wild type at all of these loci involved in DNA repair. Interestingly, as it's refined that trait, it has picked up one of these mutations that increases the mutation rate further. But it was not part of that. So there can be some aspects of evolvability-- this potential to make transitions-- which as your question implied, would allow sort of generic responses-- more recombination, higher mutation rates, perhaps some other kinds of physiological changes. But then there are these others that are kind of local evolvability. They just happened to get into a state. We think that some of the mutations, for instance, might have been they're definitely in that lineage there are some mutations affecting genes involved in the TCA cycle of which citrate is an intermediate. Their problem is getting the citrate in the cell. Well, if you put citrate into the cell without making adjustments to the TCA cycle-- by the way, this is not a result. This is a working hypothesis-- is that it sort of gums up central metabolism for an organism that's been evolving on glucose. So perhaps some of these changes in mutations in the TCA cycle may have altered their potential that then, when this good mutation comes along, they can actually make use of this additional energy. Question in the front? AUDIENCE: What happens when you throw back stationary phase? Does it change the fitness profile? RICHARD LENSKI: What about stationary phase? AUDIENCE: If you permit more time in stationary phase [INAUDIBLE]. RICHARD LENSKI: Yeah. So the question is what would happen if we varied or increased the amount of time in stationary phase. So stationary phase is the microbiologist term for this period when after they've run out of glucose, or in the case of glucose in citrate, they're sitting there starving until the next day. And some people say well, it's stationary phase because they're not growing. If you leave them starving for many, many days, they will actually start to die. But over the period of time of our experiment, we don't really see any-- so in terms of, again, fitness components-- there's just not enough time for them to see any measurable difference in their mortality. That doesn't mean, though, that there aren't physiological changes that affect that that matter. So the fact that they can go from yesterday's starvation to today's growth faster than could the ancestor, you know, suggests that they're carrying over some part of their physiological state that benefits them when they make that transition the next day. One more question. So-- AUDIENCE: My question is again about the rate of evolutionary change, and specifically, if you have thought of a way to introduce artificial extinction events to comment on [INAUDIBLE] change or [INAUDIBLE] accelerate or are somehow different after artificial extinction events versus continuous evolution, which would be for however many generations to citrate. RICHARD LENSKI: Yeah. We haven't done those kinds of experiments in this bacterial realm. We have a former student of mine, [? Gabia Didd ?] published a couple of papers in this digital realm where we sort of-- it was very easy to kind of let diversification happen and introduce extinctions and ask how those-- then those are papers in American Naturalists and I think Evolution, asking questions about how that affects the re-revolution of traits. We didn't have any direct analogy either to the citrate or to the changing mutation rates. I used to sort of-- if Stephen Jay Gould was still alive, I would've loved to have had him in the audience. I used to think that-- what I'm going to do is-- and all of you know he was a huge baseball fan. Right? And so you read about baseball. I would have loved to have a picture of my flasks and me holding a baseball, you know, getting ready to drop it over them, you know, as a like mass extinction event or whatever. But we have not done that experiment in the context that I could give a more specific answer here. But it's maybe just the last comment I would make is that this field of experimental evolution. I have this one experiment, and we have some others in the lab not nearly sold that's been very powerful, but it's been very gratifying to see so many different groups around the world starting to think and ask questions like this. And so even if we haven't done it in this system, you know, testable hypotheses are out there and exciting and keep science going. Thank you.