 Pojebno. So we can start this new session with Martina Dallbello talking about environmental responses at the organismal level to community properties part one. Good morning. Is this thing working? Can you hear me? All right, cool. Tako, izgledaj za tvoje, tudi sem bil kandljiv, tudi tvoje lekti, tako, da sem bil, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, tudi, in post-docs, in tako da. Tako. Tako. So, here we are in a... Yes, trying to connect physiology and ecology. So, here we'll try to play the ecologist of the situation and I will try to remind you why we study microbes, why we like them so much. So, basically life on earth depends on microbes. Microbes are everywhere. For example, we have them in our gut and they help us digest the food that we eat and absorb nutrients from them. But also, microbes are responsible for marine and terrestrial biogeochemistry because they are the great organic matter and cycle back important nutrients to, for example, primary producers and in doing so, they are responsible, for example, for the long-term storage of carbon. And when we think about the oceans, microbes basically produce half of the oxygen that we breathe and in doing so, across all these ecosystems, they actually regulate the production of important greenhouse gases like carbon dioxide, methane and nitrous oxide. But the point that I want to make is that rarely all these functions depend on one bug but actually depend on the complex network interactions that establish among hundreds of species organized in complex communities. So, really, it's good to care about single species but also it's important to start thinking about how all these species play together in a community. And I would say that in the past few decades with the advent of 16S amplicon sequencing with metagenomics, we have become pretty good at understanding who these microbes are. So, if I take a community in the soil and a community in the ocean, I can tell you which bugs are there, so their identity. But I think we're still struggling to understand why we see certain species in a certain part of the ocean and not in the other part of the ocean. And I think this is partly because we still need to answer one important, or at least, we're trying to answering and we're still not there yet. An important question that is how do communities change as the environment changes? And here when I think about the environment, we can think about, for example, temperature, salinity, nutrients, pH, and we know very well from what, for example, we've seen with Terry this morning and with other talks that all these environmental variables actually can act on the physiology of the single cell, of the organism itself. So, I would say that to go from the environment or changes in the environment to community structure, what we're trying to do is actually try to understand how changes at the organismal level really scale up to tell us what happens into communities. The plan for actually these three lectures is the following. Today my idea is to start with talking about the simplest communities possible. We can even decide whether we want to call them communities, but basically those composes are by two species. And how the outcomes of these interactions can change based on a concept that we call mortality burden that is telling us the relationship between how you grow and how you die. And this is actually setting, so today the plan is to set the stage to go to Thursday, where we will try to look at how the distribution of fast and slow growers change in a community, changes predictably with temperature and salinity. And I will talk specifically about marine microbes, but I think it can be extended to other systems. And finally, on Friday, the plan is to forget about temperature and salinity and go into nutrients and try to see whether we can use metabolism to try to understand how microbes assemble in communities when they grow in different resource environments. We will see where we arrive today. If we get here already, we can start moving on, because this one might be pretty long, and don't think the third one. All right, so first part. As I mentioned, we're going to look at what's the outcome of competition in the simplest communities possible in the light of mortality burden. But this mortality burden is a concept that relies on growth rates. And actually, as we have seen, growth rate is basically a pervasive concept in ecology, not just physiology. And actually, I think that ecologists have realized that growth rate is a phenotype, because so being fast or slow grower is just a possible phenotype, because being faster grower is not always the best condition possible. Because for example, when ecologists describe succession, which is the process of species colonizing, for example, an empty space, and we can think about it, for example, when a fire wipes out the forest, then ecologists describe this succession, so the arrival of different species in terms of fast and slow growth. And we know that at the beginning, the species that usually colonize are faster growing species, like grasses. And then as we proceed in time, the growth rate, the mean growth rate of the community decreases, because we will have then bushes and then the final state, usually has a lot of trees that are slow growing species. And also, well, this concept of fast versus slow growth is pervasive in microbial ecology, as well, because especially in marine environments we distinguish between copiotrofs and oligotrofs that are usually... So the copiotrofs are usually microbes that can grow very fast and can take advantage of abundant nutrients. For example, the vibrios that Terry was showing you today are actually usually classified as copiotrofs. While on the other side of the spectrum we have oligotrofs that in general show slow growth rates, very slow growth rates, and usually thrive where there are less abundant nutrients. So this is just to say that this is a very pervasive concept in ecology. We can use it, so the concept of fast versus slow growth, we can use it to describe communities. But actually the question that remains, and I think we are trying to answer, is under which conditions is it best to be a slow grower, a faster grower? Or in other words, when is the one or the other favored depending on the environmental condition? What I'm planning to show you today is that per wise co-culture experiments which means in batch, so we're not in a chemostat environment but we are in a batch culture allows us to study communities as the environment changes. And then I will show you that when you increase mortality the fast growers are actually better off, but when this outcome can be reversed if you are increasing temperature in your environment because in this case the slower growers are actually favored. So when I'm talking about these batch culture and deletion experiments, I mean experiments in which we can mix at different fractions two or more species. Here I'm talking about two species and for example, these species can be different because they have different growth rates. Measures, for example, in the media that we're using in the experiment. The other advantage of these deletion experiments is that in these flasks or whatever 96-well plates if you're doing high throughput experiments you can change the conditions in very consistent ways. You can change the concentration of nutrients. You can place them at different temperatures. You can shape them well. And the other thing you can manipulate is the dilution rate. So every day or every 48 hours depending on whatever your bags are sometimes you want to grow them for 24 sometimes it's best to grow them for 48 hours you are taking a bit of the previous day culture and you are inoculating it into fresh media. And basically this simulates so you are discarding cells and so you're basically simulating the effect of mortality which is another parameter that you can tune in these simple experiments. Well, convenient thing is that usually these two species are growing in this media and then you can actually count their abundances or relative abundances by plating or using a spectrophotometer or a flow cytometer but basically the important thing is that there are many cells or how many colony forming units you have and so you can actually get the final fraction of these communities. And I talk about final fraction because usually we do serial daily dilutions to reach a steady state. So what we are looking at is the community once it reached steady state. Since you can count abundances you can actually have an idea of the interactions and understand who is the best competitor in that specific environment. Is that clear? All right. And the other thing is that what outcomes should we expect at the end of these serial dilution passages? Well, one possible outcome is that one species out competes the other. So here as a function of time time is usually the days you are doing the dilutions or the periods. One species you see that here I am plotting the fraction of one species of the two no matter which is your starting fraction so it can be I don't know, 50-50, 5% 95% or the opposite. You always end up with this species one species out competing the other which is usually called competitive exclusion. The other outcome is that these two species can actually coexist. And here what I am showing you again is as a function of time the fraction of one species which no matter which is the initial starting fraction you end up at the end of these daily dilution passages to have a constant coexisting fractions. And finally there is another possible outcome called bistability which means that the outcome of the competition depends on the initial fraction. So who is the winner depends on the initial fraction at which you are starting experiment. For example in this case if you start with a lower fraction these species for which we are plotting the fraction is going down so the other one is winning but at a different initial fraction here is higher so what we usually see at the end of our daily dilution experiments with two bugs. Now here I am plotting two bugs. The other nice thing about these experiments is that we can easily recapitulate the dynamics of the species and the outcome of the competition with the Lotka-Volterra model. So Lotka-Volterra model has been proposed by Lotka and Volterra several I would say 100 years ago more or less and describe the dynamics of two populations. Originally it has been proposed to describe predator prey dynamics if you remember the hair and the fox what was the other one? The links, thank you. Well one predator and the other one was a prey but it has been extended also to recapitulate the outcome of competition. So here the equations are pretty simple. We have the per capita growth rate that depends on the growth rate of the species the maximum growth rate of the species and then we take into account a self inhibition term that is telling us basically how much your kins so your conspecifics are inhibiting your growth is basically how much the other competing species is inhibiting your growth and this is exemplified by a dimensionless parameter that is the inhibition coefficient. And the interpretation of this model, so the outcome and of course you have one equation for species one and one equation for species two but this model can be extended to many many species but for the moment we keep it simple. Actually so the outcome of the competition solely depends on the alpha coefficient so the inhibition coefficient. In this simple way if we plot this is a space where we have the logarithm of the alpha to one which is inhibition of species two by species one and on the x axis we have inhibition of species one we can see that we have full coexistence when both coefficients are smaller than one we have by stability with both coefficients are larger than one and you can have either dominance of one species or exclusion or the same species whether one of the two coefficients is one is larger than one and one is smaller than one and these are all analytical solutions so for this model actually recapitulates perfectly all the possible outcomes that we see in species in co-culture experiments so the question that then you can ask is okay if I'm thinking about the environment can what changes due to the environment can happen and can be actually recapitulated in this model for example one possible thing that you can think about is well maybe there is environmental deterioration and just to give you an idea a common form of environmental deterioration is an increased mortality and let's think about overfishing if you are you have a population of fish and you are going there with your big boats you are removing a lot of fishes and what happens is that you are applying let's say a mortality that is global so everyone is more or less so on you the mortality that is arriving on you is basically similar across those species and actually the outcome we know for example for several communities in Teranova banks where many species is being wiped out is that the outcome of this increased mortality is a change in community structure so one question that we can try to ask in these simple systems is how do changes in mortality affects the outcome of the species competition and the good thing is that in these simple experiments as if you remember before we have a simple way of manipulating mortality so when we do these daily dilutions we are actually discarding a fraction of cells and how much are diluting yes in this kind of mortality you are removing the population from the reactor or the sea but there is another kind of mortality where you kill them and you release nutrients which would give a very different outcome I agree this is mortality when you just wipe out the population and there is nothing for you to gain if you are the other species and also in these systems especially if you are growing them every 24 hours actually you rarely see people dying so the nutrients that are actually exposed to are usually nutrients that we provide or that are produced by cross feeding but I don't want to get into this now during the growth rate whether it's exponential or in steady state over many cycles so with the Lodka Voltero we are actually looking at steady state dynamics not at the transient it's actually you arrive at the steady state that's why the serial dilution is actually to be sure that you're not in a transient state but you're looking at the final state of the community that is stable you kill most of them every time but basically every time you reach a stable population and so you ensure that who is actually surviving so that survives who is actually surviving the mortality so the dilution rate let's say if you grow less than your mortality you're deterministically wiped out if you grow more than the mortality then you can survive but then the outcome so who is winning actually depends on several things most of the competition coefficients but the competition coefficients depends on your growth rate and how you actually are competing against mortality which is where I'm going basically does it? ok as I was saying so basically in these experiments you can simulate this global mortality where you're not killing cells you're just eliminating them from the population and you can survive in this system only if your growth rate is actually higher than the mortality rate we can go back to this and actually one question is that usually when you are competing these pieces you don't have a species that is growing faster than the other in a certain condition that you are analyzing so it's pretty easy to add mortality rate to not cover terror models and specifically this mortality that we're adding that wants to simulate what we're doing in the experiments is this constant so it's a constant fraction of cells at each dilution rate and you can do some reparamentalization of the model to absorb mortality I don't think it's super important to go into the math but basically what this reparamentalization allows you to do is to write the competition coefficients here I wrote the competition coefficients so this is the inhibition of the slow species by the fast pieces which is just a function of the mortality and the growth rate so basically how you are surviving in this system depends on this ratio between mortality and growth rates and I call this actually we can call this ratio a mortality burden which is telling us basically how good you are at out competing mortality which is fixed or can change and depends on how fast you grow I think it should be pretty intuitive so if you grow fast enough to overcome mortality you can survive but then when you actually look at the competition so the competition coefficients how much you are good at inhibiting the other species depends exactly on how good you are at competing mortality you would just keep the fractions the same well once you reach that state then it becomes the same if you just take a random sample and put it in a new medium you expect the same fractions before and after well you expect so you start with your populations they reach certain fractions and then you dilute and then so you start with these initial fractions so the idea of the dilution is actually to end up with a stable fraction so once you actually don't change the fractions it means that the population is stable but at the beginning you can have changes because they might grow in different ways and you are when you are removing them they arrive at a certain fraction and then you are diluting at that fraction ok so doesn't have mortality to have a term that's multiplied by n? yeah yeah yeah so the n is dividing everything oh sorry ok so n is actually multiplying r and multiplying n ok cool the idea is that these competition coefficients only depend on growth rate and mortality and the thing is that we can actually if we go back to our previous plot we can see that we can still plot in the same way where actually we can go in this section where this piece is one slower grower is winning the fast grower is winning and here we have by stability and coexistence actually in this model the cool thing is that it is possible for a slow grower to out compete a fast grower when mortality is sufficiently low then as we increase mortality what we are doing I think it's a good way to visualize this is that you are moving through this space by 45 degrees so let's say if you start here and you are and here the slow grower is winning and we are at low mortality rate if we increase mortality we can either cross a coexistence region and then end up with a faster growing winning or if you are for example here with here I mean the relationship between the two competing coefficients you can still have the slower growing winning cross a region of by stability and then end up with the fast grower winning so this is how the growth rate is affecting sorry the mortality rate is affecting these ratios and affecting therefore the relationship between competition coefficients I see many perplexed faces ok let's do this let's look at experiment no ten so there's a difference between the growth rate that's realized that's sort of a net growth rate of all of these processes but you're thinking about the one period during one period if you are actually making to the second step it means that you are growing faster than the other one at certain point but over longer periods you can be a slow grower on average so across different mortality rates but at low dilutions these ratios in these ways ensure that you're still out competing the fast grower but as soon as you are increasing mortality the fast grower becomes better at out competing mortality itself and so becomes a better competitor I think yeah that makes sense and I think that I mean the math is undeniably the same the confusion for me is that it says slow grower wins because there's this intuition that to win at some instant in time you need to have a higher growth rate and I think there's two growth rates here there's like an instantaneous growth rate that's basically defined by this is your maximum growth rate so the thing that is defining you as the slow grower, the fast grower is your maximum growth rate over the cycle so when you're actually growing you can there are differences in growth rate but let's say if I'm looking at if I'm measuring growth rate in this condition in this growth media you can be always lower than me, I can always be faster than you with my maximum growth rate but as I change mortality even if this relationship doesn't change you can out compete me at low mortality and I can out compete you at low mortality because at some instant in point the effective growth rate that you have they are both surviving over the cycle, yes let's figure this out in the break thank you when you say increase mortality rate that also can be done with diluting we are doing it with dilution so the more I dilute the more the fast grower wins that's the idea can you say why it is not a problem that there's one constant mortality rate instead of this serial dilution that you're actually doing the mortality rate is actually exactly exemplifying these dilutions it's constant because you're always removing the same fraction of cells then the fraction of cells that you are removing can have different fractions of the two species so this is a chemostat it's a chemostat model but actually if you do like the batch culture model it doesn't really change so the thing here is that you can approximate a chemostat model to a batch culture or at least that's our I have sort of more of a comment than a question because from the previous questions I can see that people are confused about the slower species winning over the fast species but the reason why it wins is because there is a strong suppression of the growth of the fast species by the slower species so in other words those interaction terms is what makes the slower species to win if they don't interact of course the fast species will always overtake if you set your alphas to zero then no question it affects the competition coefficient so actually this is the way to see how this actually competition coefficient changes and so in this region one so the competition coefficient for the fast grower is less than one and the other one is larger than one and so that is why the fast grower is winning in terms of numbers how the dilution rate or mortality rate in your experiments compared to the growth rate of the fast and the grower slowgrowing species because from that expression if the dilution rate is the same rate as the slowgrowing species the interaction coefficient should explode so let's say if the slowgrowing species has a growth rate of 0.3 per hour and the fast one per hour can I show you the data experiment alright so for example this is growth rate of two species this is enterobactererogenous formerly known enterobactererogenous, I think now is glepsiella possibly you can see that it's faster compared to these other pseudomonas species so it reaches saturation before the other species and now we are doing an experiment with the dilution factor of 10 which means that I am transferring one tenth of the of the previous day culture into the new culture which is a quite low dilution rate and you can see that over the course of time here I am plotting the fraction of the fast grower so the red guy you can see that over the course of time the dilution passages at the end the fast grower is actually outcompeted and the slow grower is winning so this is how ok maybe I don't have the actual number for the growth rates but I can take them but so the dilution factor is is 10 so mortality would be 0.1 yes so because it's not in rich media this is in minimum so are you putting rich or not has to do with the speed but also probably it has to do with volume has to do with how much how much nutrient you are putting you are putting a very low amount yeah it's not high I think it's 0.1 or 0.2 does the result change if you I think yeah I mean if you increase the concentration they would grow to higher yeah it's it's not I agree but it's a minimal media and with not too much carbon in it I don't exactly might be like 0.1 or 0.2 percent per volume no the other two I think it's either it should be M9 or another medium so you're not limited by nitrogen and phosphorus but they can be carbon limited even though they do you understand in this case how fast growers well it's we think it's because of this mechanism we were thinking about before that here the slow grower is actually inhibiting more because it's able to grow and so it depends on the relationship I will show you within the lack of athera as Leonardo was commenting you would need to propose slow growers and fast growers but do you know? no I don't know exactly what they are doing there could be many other things you get exponential growth during the growth period? I think so this is a single species these are single species during co-culture I don't know I can check but at this very low nutrient level I can check it shouldn't be that much more but we can check I don't remember now alright so yeah sorry one more quick question so I totally understood this whole thing when Sergei explained it but that means a corollary of that is that this should not be generic because in some cases some species the slow growers the fast growers but I think the reason why we are not going to mechanisms because it doesn't necessarily have to inhibit it can be better at consuming the resource but all these things are contingencies this may work 50% they are contingencies but they happen quite often which means even though you don't know you don't exactly know the mechanism except that you are growing you are growing at a certain rate you are either growing faster you have this relation between mortality and growth rate the mechanism for which they are inhibiting each other we don't really care about it we are not asking how but it's sufficiently I would say generic that no matter what they are doing it happens it's not about the exact mechanism it sounds like it could sometimes maybe but I'm saying it's just it's an interaction and I'm not putting a label on the type of interaction I don't know the type of interaction I just want to comment here that the interaction must be necessarily not through resource competition because if you are there competing for a single resource no it's not a single resource it's not, it's a rich medium no it's not a rich medium you can do it in two resources you can do it in one resource and it doesn't change so then the inhibition must be through pH or whatever other mechanism it's buffered then it's a bit of a mystery because if you kind of look at your growth curves it should be independent of the dilution factor the fact that the red one has faster exponential growth that should give you a dominant effect of the red one if it's only talking about competition for resources so something else must be going on I don't know how they're I think we can move forward so this experiment done with five species and they showed that basically predictably you can always establish when you do pairs one species that is slower growing slower growing compared to the other one and as you increase the dilution rate here I'm going you can for example these two species that is these interbactererogenes and sedomonas veroni you can cross the coexistence region and you can end up with the sedomonas veroni winning as you increase mortality but actually you can also have the other possibility here we have again sedomonas veroni but it's slower growing compared to other species that is sedomonas citronellalis and here you start always with sedomonas veroni winning at lower dilution rate but you cross the by stability region in which the winner depends on the initial fraction and finally you can end up with the slower faster growing what I'm saying is that you cannot believe the model you cannot believe the mechanism this is the data yes one second good question good question I think that you can reproduce these models if you actually look at growth with lack time I think Yakupo can tell you more about this what do I do what do I do those are different species what's so what's the dilution rate so you go like 10, 100,000 or maybe so the dilution dilution factor yes no you don't change from time you do one run at 10 one run at 100 and one run at 1000 yes no, but you're always 24 hours so there's a lot to do there's a lot to do of course so I'm an experimentalist, I believe the data then we use a phenomenological model to try to understand what's going on but I think that at the end of the day bothers me that these predictions, whatever you do mortality you do temperature, you do salinity you always get the same things maybe the model is completely wrong but it's helping us actually making predictions that then we can test and when we test them they turn out to be true I don't care about the models honestly single species so this is not my data this was trying to set up the stage to arrive at my data at a certain point but these are all interesting things that we don't know actually and so I think that hopefully for the next 7 years which is my tenure anyway I'm kidding again I think the main takeaway here is that the model is a phenomenological model so it doesn't have any mechanism attached it's just trying to understand the dynamics and actually what we see is that to me the most important thing is that you can modulate the outcome of the competition by modulating mortality which is and by modulating the environment let's say this so the question is can we actually predict something harder so can we actually predict how temperature can affect competition again temperature is so what time is it how much do I have if we count the number of questions you have like 45 minutes no I'm joking I think we can go for 10 minutes ok cool then it's what I need alright so temperature we've seen is everyone knows is a cardinal variable of life and also especially microbial life because bugs are not able to regulate their internal temperature and this is taken from a basic microbiology book and it shows that each bacterium has a specific growth curve as a function of temperature and it has a minimum temperature at which it can grow below that temperature you can have membrane gelling so some transporters cannot work anymore and so the species cannot grow then you have growth rate is increasing rapidly because temperature is actually increasing the rate of enzymatic reactions and then at a certain point you reach an optimum that you can grow species at the optimum temperature and then above the optimum then you start having all these problems protein starts to denaturate and then you have the collapse of the membranes you can have lysis so bad, bad, bad you die the other interesting thing is that we have microbes actually can literally explore the range of temperatures that we have on our planet because they can grow at very different temperatures meaning that they have very different optima we've seen this morning that for example vibro nitrogen likes 37 degrees bacillus likes a higher temperature but these bugs are actually all mesophiles they like this range of temperature between 10 and 50 degrees but then you can go there are some species that they like the cold and others that can reach they grow best at 100 degrees celsius so they thrive in these crazy environments that for example thermal springs so let's say we stay in the realm of mesophiles that is what we can grow in the lab and we can ask at a certain temperature which species is going to benefit in our experiments here I'm showing you the result of a co-culture experiment same experiments as before we mix two species we mix them at different fractions the species for if you are interested are two species isolated from soil they are pretty easy to grow they like everything we grow them in these minimal medium plus carbon sources and we see that at this intermediate temperature they coexist they reach this stable fraction they like each other at this we can actually measure the maximum growth rate at these two temperatures and we see that the red guy is growing faster than the blue guy and actually there is almost a two fold difference in their maximum growth rate despite that they coexist so the question is and here I'm plotting the fraction of the blue guy and what we will see in the next two plots is the fraction of the blue guy what happens if we increase the temperature well very predictably because we've just seen the curves before both species increase their growth rate so both the red guy and the blue guy going from 16 degrees to 30 degrees see an increase in their growth rate and the red guy is actually increasing its growth rate much more so the the increasing growth rate is usually higher compared the proportion the difference between the two growth rate increases as we increase temperature and now you can ask if I compete these two species at this temperature at which I measure growth rate what's the outcome of the competition at lower temperature is out competed by the fast grower but I would say you might say surprisingly the slow grower is out competed the fast grower at the highest temperature even though there is still a difference in temperature in growth rate and this difference in growth rate is actually favoring the fast grower then I'm trying to get the approach that we were telling before so what can happen when we increase temperature for example higher temperatures can activate the antibiotic production in some species or we've seen for example there was a poster yesterday I think it was Benjamin that you can actually induce thermalizes by a phase if you increase temperature so all these possible things can happen when you increase temperature actually one thing that we can do in our simple phenomenological model that we can believe or not believe is we can allow growth rate to change as a function of temperature this is the same model that we've seen before we still have mortality because we're still trying to understand what's going on in these daily dilution experiments in which I'm imposing a fixed mortality rate this thing is not changing across experiments the only thing that's changing is the growth rate and we can try to model implement growth rates in this model as a function of temperature and here I'm just reminding you what we can do we can always do this parameterization in which the competition coefficient only depends on the ratio between mortality and growth rates that's a good point but I think it can so if you're actually increasing growth rates so what I'm saying is that your inhibition still depends on your growth rate and if your growth rate is changing with temperature also your inhibition factor which is the whole point of this thing is also changing with temperature yes they are both mesofiles it's kind of the same no no no no no so this is an important point we are just about to arrive there we are within the thermal limits below the optimum the thing is just that this is how in the model you can change the growth rates so this is how increase in temperature and this is how for example the maximum growth rate change and one important thing that I want to show you just to make you understand a bit more this concept is that this ratio depends on mortality which is fixed and growth rates that are changing this is how it changes for the fast grower and the slow grower as temperature increases Simon measured growth rate at different temperatures and they were not in the collapsing phase so actually the thing that I wanted I can make really quickly probably is that you can model these dependence on temperatures in two ways you can have maybe I can go there, this is a renews and you are assuming that there is linear dependence between growth rate and the inverse of temperature and so basically you can say that the growth rate is a function of again, it's an exponential function of the activation energy, the Boltzmann constant and the temperature in K and the way you can multiply each growth rate by a certain factor to distinguish between a fast and a slow grower but you can actually do it in a different way so in this paper that is published actually Simon and Claire used a different model the Ratkowski model that is actually slightly more complex but fits the temperature curves much better and here growth rate is a function of temperature depends on the slope of the square root of the growth rate versus temperature and you take into account this minimal growth temperature but the point is no matter the way you are including the dependence on temperature of growth rate let's say the outcome of this model shows that but here just to remind you what the increased mortality rate was doing so shifting the outcome from this quadrant where the slow growers win to this quadrant where the fast growers wins temperature is doing the opposite so shifting back the outcome of the competition towards the slower growers winning so if you are at a certain dilution rate here you keep the dilution rate fixed but you increase the temperature the outcome of the competition is that the slower growers is going to win and this is how this mortality burden is affecting the competitive coefficients and so in this quadrant here the slow grower is inhibiting more the fast grower ok so how often does it happen in our per-wiseco cultures? quite often before I showed you the results of two pairs of one pair but in this heat map I am plotting the fraction of the slow grower and you see that if you move from 16 degrees the temperature that Simon and Claire looked at you generally see that the relative abundance of the slow grower increases as temperature increases so it's a pretty widespread phenomenon so I think I can just wrap up here and well maybe I convinced you that per-wiseco culture experiments are a nice playground to look at how the environment can change the outcome of competition and I hope that at least the data convinced you that if you increase mortality your favor in fast growing species this is all relative and instead increase temperature can reverse this effect and actually favor slow growers the plan is that I'll show you that for temperature this is actually possibly true in natural microbiomes and for today I guess we can stop here