15:05:39 that I am telling you. Probably I want to have the time to tell you both, but at least you know what I'm talking about. 15:05:47 And. 15:05:49 Okay. 15:05:51 So, I want, I wanted to start with a collection of beautiful pictures of places where you can find my props, basically to distract you from the fact that I'm gonna tell you something that has been repeated over and over, I guess over the past three weeks. 15:06:07 And that is that microbes live everywhere from our skin to our gas from the other several events to the hot springs and basically microbes don't live in isolation but form these complex communities, which are essential for the health and function of virtually 15:06:32 all these eco systems. So basically this is why we should care about microbial communities. But there are basic question about microbial communities that are still not completely answered. 15:06:44 And for example, sorry. 15:06:45 We still don't know what are the determinants of microbial community diversity. So, in other words, what are the factors that allow many species that way too quick system at the same time in the same community, and the other things that we don't know 15:07:00 is what are the mechanisms behind changes in community structure that we observe across in time and space across selected to do spatial and temporal regions way I like to think about these questions is through a metabolic lens, and I'm not the only one 15:07:26 here there are a lot of people that think about these things. But, basically. So if we think about organisms as energy processors, then microbes are actually microbes are the best because the link between metabolism and quality is very stronger. 15:07:42 So in other words, If we think about the how the environment that can shape the metabolism, we can think about resources we can think about temperature. 15:08:03 byproducts can shape and modify it environments that surrounds them. So this is a very strong coupling. And this I think allows to link ecological outcomes to physical processes across different organizations, organizational skills and so from cell sales 15:08:19 to communities. 15:08:23 And so today I'm going to tell you something about so important environmental drivers that can act on microbial metabolism and the resources in temperature. 15:08:35 And so, the first story is about resources, and it's about this a relationship between the available resources, and the community diversity, and I will also try to tell you something about some follow up questions that stemmed from this project. 15:08:57 In the time today is about in this moment. And if there is time. I doubt it. 15:09:15 I will show you. I will tell you something about the different story about how temperature can be the driver of changes in composition of marine bacteria communities across seasons latitudes and depth. 15:09:18 So, the first story. So this is work that I've done in the burlap bag. 15:09:28 Very recently. 15:09:34 And this is the work that I've done in collaboration with actually that is here and everyone knows, and young so clean. 15:09:40 And again, it's about the relationship between diversity, and the available resources in the environment. 15:09:46 So among all the different drivers that can explain diversity microbial communities, and there is for example the space of structural environment that can be predator free dynamics the pages, and many other processes to say the interaction for resources, 15:10:09 received a lot of attention, both theoretically and experimentally. And I guess that everyone is. 15:10:14 I have problems are with everyone here I guess he's aware of the fact that several models are based on the principle of competitive exclusion, and based on this principle, if I provide a single source of carbon, for example glucose, then the equilibrium, 15:10:32 I will have only one species surviving. 15:10:37 And the species that survives is also the one that has the fastest growth rate on this resource. 15:10:45 That's when what, when people did experiments about that they actually found that when a single source of source of carbon is provided. Then, multiple species can actually quick system, and this has been shown for a glucose but those are some amino acids. 15:11:04 And as a red been said multiple times student is the program. One of the mechanism that can be invoked to explain species persistence is course feeding, which is the phenomenon by which the metabolic byproducts, that are produced during during group bacterial 15:11:23 growth can leak in the environment and becomes food for other community members. 15:11:29 What we don't know is what happens when instead of having just one resource. I provide multiple resources. So basically, we still don't know what is the link between fermented diversity and the number of times, and type of resources that are provided 15:11:44 in the medium. 15:11:46 the medium. So, just to help you follow the story. I'm going to tell you what I'm going to show you physically I'm going to show you that several resources when provided a single source of carbon and support the rich multi species communities. 15:12:02 But when these resources are combined community diversity increases only modestly within with additional resources. 15:12:12 And we. 15:12:14 This apparently contrast the results can be reconciled. When the CrossFit in network that is seated by the supplied resources is taking into account. 15:12:27 So, we tackle this question about the relationship between resources, and committed diversity within experiment. 15:12:39 With an experiment. And I started with this soil sample, and I literally went outside the, the building at MIT. And I went to the grassy knoll and this is in front of our building with an apple corer, and I scoop the Beatles soil, I came back to the lab, 15:12:55 And I scoop the Beatles soil, I came back to the lab, I am mechanically detached, the cells that are attached to the sofa articles by shaking this sample. 15:13:10 And I obtained this microbial suspension that later on, I learned that contains more than 700 the species. 15:13:14 And I use this microbial suspension to inoculate the 75 different resource environment and to be these resource environments and I started with a pool of 16 carbon source. 15:13:28 So in this pool of resources. 15:13:31 I included some mono saccharine like glucose and fructose, we had also some saccharine like sucrose. We had a collection of organic acids. 15:13:46 Sugar articles, we have, we had also to policymakers cellulose and start. We know we had also these either oxy Brolin, I'd say probably music is similar to an amino acid is a compound that is commonly found in the soil, soil because is, it is released 15:13:55 by the roots of the plans. 15:14:07 From what I know and I use these. 15:14:09 Each of these carbon sources, could be provided as the single sources of carbon, or in combination of 248 10 and I also had all the 16 resources together. 15:14:24 And the carbon concentration across all these conditions, was kept Constanta at 100% with per volume. 15:14:32 I want to tell you another thing about how we did this experiment that is, we use the sort of tournament structure to build the different resource environments. 15:14:41 In this way, we could basically compare the communities that were brought in a combination of resources with the communities grown in the constituent pairs or singles resources. 15:14:55 Is this clear 15:15:15 like what was your nitrogen tours and so on. 15:15:25 competitive exclusion principles, a priori only says that no two organisms can be limited by the same resource, not the same as by the same reasons that you add, so it might be limited by oxygen. 15:15:34 Others might be limited by a vitamin and so on. How did you incorporate those Yeah, so the the carbon source was the 2am nine medium that has sources of phosphorus and nitrogen is the phosphate buffer and there is a ammonium chloride. 15:15:49 And I added vitamins as well. Just because for voice can be sufficient to make sure that those will not be the limiting we should I think so. Okay, that's my guess. 15:15:59 And they were will meet so they were shaking. So this is the place. So these are 96 or plates, deep 9264 plates, but with the is is still on top of that allows exchange of oxygen, and there is enough volume so this is wondering, 10 microliter in 500 milliliters 15:16:18 deep well, which means there is sufficient space to allow the exchange of oxygen. They were shaking. 15:16:25 They were 30 degrees. 15:16:30 And the youngsters. 15:16:31 Okay, question. Yeah, thanks. 15:16:35 I kind of have a follow up along the same lines, how about amino acids, I know that you cannot provide them because they can be used as a carbon source, but uh, what if you have oxygen drops in certain amino acids, they could be the limited. 15:16:50 So, if this is the case, I might. 15:16:54 But the thing is if I started adding to the problem with amino acids. Not a problem but it's something that you have to think about that you're also modifying the source of nitrogen. 15:17:03 So, and in this case I was not much time not to modify the source of either the source of nitrogen. 15:17:10 And what about a serial, you did a serial daters Well, yeah. 15:17:18 So, yes we did the daily the Russian protocol. So, and it was a 25 1024 hour dilution protocol, which means that at the end of each 24 hour growth cycle I was taking a small amount of the previous day culture and integrating it into fresh media, and I 15:17:35 did this for seven days. 15:17:39 And, which is enough time. So we checked after was but it's enough time for the committee with them, and they serve and I also the days before I expected a DNA, and physically stand for sequencing for 16 hours. 15:17:58 So you just said that amino acids are problem but I think hydroxyl proline is an amino acid yes, no, no. Did you do anything differently in that case to know I chose that I that's a problem because it's quite abundant in the soil, and I thought it was 15:18:12 interesting to see how it behaves. But I agree that this changes the nitrogen, and it's similar to me. Awesome. Yes. And then maybe you can just comment first because I you know in my head all of these carbon sources are kind of in categories and roughly 15:18:26 in the ones that you drew them, you have any expectations going in about, you know, what the behavior would be like in these different carbon sources. 15:18:34 Yeah, I honestly I was thinking that maybe the complexity or basically the molecular weight would give you an idea of how it needs to diversity single resources would behave. 15:18:46 but I will show you later that didn't work. 15:18:51 But yeah disposition was I'm, I'm trying to include the as much as I can, different resources with different complex, and indeed this wasn't one carbon to a sorry to one sugar two sugars party saccharine the yard calls to try to have a different mix 15:19:23 You have a tournament structure, I don't know what exactly you mean by this so could you explain this a little better. Yes. So let's say that, okay this is the this is the order with which I provide this is. 15:19:31 Yes, knowing that I am. Nope. 15:19:37 Yes, then expect. 15:19:39 Yeah. Okay, so let's say that this is the economics my course. 15:19:45 So that's it. This is the, the order with which I provide the carbon sources. 15:20:01 And then the combination of eight will be together we say that if you met. 15:20:04 Then I do, I pair glucose and fructose and Silas and mangoes. And then the for the combination of four will be glucose, fructose silos. 15:20:07 Ronnie, is the order. And I switch the order three times. 15:20:10 is the order. And as which the order three times. So you order the carbon sources, you pair them. 15:20:14 And I did three times. So this is what I mean we turn on structure. 15:20:24 Right. So, this order, we want to force and then you read shuffle everything randomly repeated. 15:20:33 Thank you. 15:20:36 I'm great. So yes, this is this tournament structure was talking about and it's handy later if you want to compare, let's say, the pair of the resources with the constituent symbols and so on so forth. 15:20:58 Great. So, let's start with the results I'm gonna start showing you the communities that were grown in single resources. 15:20:58 And so, as I was saying before, as you pointed out, I order them based on their complexity on the y axis on the x axis I have the richness measure there's the number of A's. 15:21:13 And so here you can see that all the power of the resources I provided the single source of carbon were able to support that multi species communities, with an average richness of about 22 species. 15:21:28 Yes, sure. Maybe you said this is it, equal concentrations of each other. 15:21:34 carbon atoms or No. 15:21:38 Okay. 15:21:39 This point 1% with the volume. 15:21:44 I'm so Ivan Introduce yourself 22 species but the most striking thing is that there is a huge body ability around this mean with some resources like citrate and Fumarate and proline a supporting just those little species, while other resources like settlers 15:22:06 supporting more than 40 species. And what about how many generations total or women to calculate calculation. 15:22:16 60, something like this 16 generations. So, you did multiple cycles right so yeah I did seven cycles, and each cycles 16 generations. Okay so that is it is it is to each cycle. 15:22:29 So the other growth rates very different. I mean, I assume they're very different way so, but you control for the number of generations. 15:22:34 You don't count. 15:22:38 What was the dilution rate, because the 30 x seven generations so 39 some 50, yes yeah I didn't 160 15:23:02 Okay, I don't know the growth of the bacteria because these are so common, so I don't know exactly which species what kind of grocery does. Although just community growth rate I don't know the total biomass is constant. 15:23:07 Is it is it was kind of, it was similar. So, so I don't have this plot but basically sorry I didn't understand. But I have a plot somewhere but it's so basically we found that the biomass across single resource, then also across all the other combinations 15:23:23 were kind of similar. So it was not really affected by the type of resources. 15:23:36 Okay, so um yeah I was measuring the variability across this huge variation around the the average richness, with some resources supporting few species and other resources like stellar supporting more than 40 species and. 15:23:52 Well, the first thing is okay what can be causing these first ecosystem of so many species, and the most likely explanation is the proceedings. 15:24:06 I don't know why. 15:24:08 Anyway, and but. And the other thing is, okay what is causing these variability. And as Avi has guests before. Well, the first the first way, the first thing that came to our mind was trying to see whether the molecular weight could predict the changes 15:24:27 and richness in single resources, but as you can see this is not really the case because the director is actually had the addition is it was lower than some more side of a question is you, you said the most likely explanation is cross fitting. 15:24:45 Have you excluded the even more likely explanation that your communities have not yet, actually reached that studies yes I know there is this stage, so I asked you had like a leveling of normal growth. 15:24:59 Yeah, so the growth is, is constant, so like the 67 the biomass is that was the same. 15:25:04 And then we also check the abundance of the, the relative abundance and the communities that calibrate after four days. 15:25:23 seven days. Good timing to reach them. So, yes. Thanks, 15:25:29 a fraction of the bar. 15:25:30 So the. 15:25:33 For example, yes so what fraction. What fraction abundance is the largest is the highest for one species so the the biomass. 15:25:46 Yeah biomass is 16 s data and study course, it depends on the resource, so some, like the, these, these citrate perfumery that you have a very abundant is V, but for example cellulose the community is much more even. 15:26:01 So it depends on the resource so. 15:26:06 Okay, so, um, well molecular weight is not really working. So we started thinking if it is course feeding, is there a way to try to estimate the, the extent of breastfeeding, or in other words how many metabolites are available in their environment, depending 15:26:23 on the resource and in genetic in symbol of microbes. 15:26:28 So what I'm. 15:26:29 What I have here is a simplified map of the microbial metabolism and I guess the, you can recognize the quality is this the tissue a cycle, and this is very simplified that I'm missing a lot of steps but it's a way to try to explain what we did. 15:26:47 So, and you can see that I liked it with different colors the position in which the resources that we provided in the media, and there in the, in the, in the metabolic pathways. 15:27:00 And what we did was we used this map reactions which correspond to a genetic and symbol of microbes. And we asked if we provide one resource and some currency molecules that can be a teepee can be can be water, um, any intermediate Can I generate. 15:27:18 In other words, how many rock productions Can I open and how many kilometers, can I generate, and the disco scope of function analysis and there are very well here. 15:27:27 But basically, we were able to estimate a number of metabolites British resources we provided in the media. And when we plotted against the richness of the communities we found that there is a strong correlation between the number of metabolize and the 15:27:41 the richness and disclaimer before question this is a rough estimate. We are not claiming that is correct. Because there are projects get can stay that stay inside the cell that are not leaked outside but at least this give us a decent way to predict 15:28:01 the regions to this confuses me a little bit because basically everything on this list. 15:28:08 I can grow Nikolai on as the sole source of carbon so that leads me to believe that presumably I can make nearly everything in its biomass function, out of that thing, meaning like you know everything that goes into coolers biomass can be made out of 15:28:22 fructose and certainly starch and certainly out of glucose etc. So how should I understand this result that you can get different numbers of metabolites like What are you searching over to get the result and. 15:28:39 Yeah. Was it again very it off and Wendy. 15:28:39 Okay, well I yes but basically we it's a it's a genetic and somewhat of Microsoft on the equalizer, but it is that simply. We're just counting how many intermediates were produced, which means also intermediates it might not be by three main. 15:28:56 Yes catabolic intermediate. 15:29:00 Okay. Can I ask a question, if I throw out cellulose in Scituate in Fumarate, I'm not sure I see anything. 15:29:08 I mean, because we did something you know you know we have lots of. 15:29:20 Watch has a lot I mean from like five years ago we never saw anything like this when we counted and I would argue that it's actually just three data points that are giving you even any trend here, so you don't believe that you can get. 15:29:25 Throughout the cellulose throw a few right and citrate and it's not clear you would get anything. 15:29:32 I don't I don't fully understand. Do you mean that you don't you don't get any intermediate. 15:29:37 Yeah, that's what I mean i think i think it's being dominated by three points. That's all I'm saying. So not so good is that if you give STI three then you can generate the. 15:29:56 Sorry disappear some correlation. I mean, if you yeah okay I understand that you use you eliminate 10 rows and fewer than cited then I think that, okay. 15:30:08 The correlation is not great. I haven't tried removing those though. 15:30:12 It was convincing. 15:30:17 Those likely like point seven something like this. 15:30:22 Sorry I couldn't understand. 15:30:25 Okay. Um, well, can this was a good proxy for us, or at least we believed it. 15:30:33 And, Yeah, so, and what else. 15:30:38 The other thing is that when we look when we saw these results with these large number of species we we were kind of expecting that in the moment we add more resources, we would see a rapid increase in diversity. 15:30:53 And so we started thinking in the ways. Oh, sorry, sorry, I forgot one thing to say. I got confused. Okay. One thing that maybe I wanted to say before going. 15:31:05 Moving forward, is that these resources that actually were supported the less niche communities. 15:31:13 If you look at their position in the, in the map, they all belong to the TC cycle or enter in the pathways at the level of literacy cycle. And so we can these resources are usually identify this energetic resources because they, they get the reaction 15:31:44 the opposite way of the, of the economic policies, by the the sugars, usually are recognized as democratic resource, because they trigger the fluxes in the direction of regular colleagues. And the thing is, it seems that the far away, far away you are 15:31:48 from the most far away you are from the further away you are from the center metabolic pathway. The more diversity sustain issue believe it, then yes so the other thing is that we see these large diversity in single communities. 15:32:04 And we thought well, if we provide more resources probably diversity would increase at a very rapid pace. 15:32:11 And so the thing that we try to do next was trying to predict the diversity in two resources, given the diversity in single resources. 15:32:23 Okay, so, yeah, I'm just brought in the richness dust on the y axis. 15:32:40 And let's say let's take for example, to two communities and the community is growing glucose and the communities were on you know the Roxy for all in one way to predict the richness in the combination of glucose and crawling is thinking that Okay, I 15:32:47 I have some species that surviving glucose, they will survive in the combination of Brooks and Pauline, I have some species surviving probably, they will probably survival so in that combination. 15:32:57 So basically the richness in glucose and Colleen wouldn't be the union of the two communities. And this position work pretty well. The another paper that was published recently, in which they looked at the composition of the community is grown in resources 15:33:13 that are usually released by micro-algae. 15:33:18 You can also think that maybe these are too many species, and maybe the best way to predict the richness in glucose enthralling is the maximum richness of the tourist horses, but maybe they're the limited they're limited by something. 15:33:30 But actually, when we measured the richness in glucose and proline. We saw fewer species compared to the value that we predicted. And actually, the best way to predict the value that is about 17 species, was to take the average of the two communities. 15:33:48 And then you can expect these these the rule, or this is the, Is this the exception. 15:33:53 And actually, now I'm plotting, all the combination of two resources. 15:33:59 And you can see and on the again. 15:34:03 On the left we have the single resources on the right that we have the best of resources and these lines connect the consistency and goes to the correspondent there. 15:34:15 And you can see, even though this plot is a bit dense that the majority will line converge towards the middle. So basically, the best way to predict the richness in terms of resources was sticking the average of the research and symbols. 15:34:30 And you can you can actually check the prediction so we compute all the prediction based on the Union and the maximum, and we compare them with the mean, and indeed the, the, the mean is actually the best way to predict that it turns into resources. 15:34:46 This also means that if you, if you when you are the water source of the media, then you don't really see an increase in diversity. 15:34:56 And these are in. 15:34:59 Again, these are already telling us something, telling us that probably when we increase the when we provide multiple carbon sources, then we couldn't see a huge increase in diverse, it but the striking thing here is that increase is really linear. 15:35:14 So basically, the diversity increase slowly as the constant rate of about one two species for each new resource that we are the, which is again similar to the competitive exclusion principle which says that you cannot have more resources than the limited 15:35:30 more species than the limited resources but still, it has a huge upset that is provided by the richness in single resource. 15:35:39 Quick question about how you're defining the existence of the species, is it a lower bound on the fractional abundance of read for know what this is without the lower but no What about that. 15:35:49 But then, like one or two you for like one observation of a species you define it as being there. Yes, but that then doesn't change if you do a cut off. 15:35:59 So if you cut the biomass at the lower bound. The train is exactly the same 15:36:08 as again follow up on this question, you, you said that you test different cycles. How do you make sure that the this particular law abundance species is not on its way out of the community that it actually reached a steady state, in a dilution cycle. 15:36:26 Okay, these I didn't properly check. So, I, but the thing is we tried different this, when we, when we saw that trend was basta to the exclusion of low abandoned species we can have class. 15:36:40 a cycle before or actually we have the full time city so right then you can for instance, argue that if your species is exponentially decreasing just you didn't have time yet to completely eliminated. 15:36:59 Then abundance at the last cycle would be less than the abundance and the previous cycle. I think that that will be the trend on this, on the same day before cycle was also the same. 15:37:11 But yeah, something that, but it's, I agree that is something 15:37:20 that I'm curious. So, is there a dose dependence for individual carbon so let's say you take glucose. Glucose plus glucose is that different. You know twice the amount of glucose will be a different relative to, you know, half the amount of glucose. 15:37:36 Yet, twice the amount of proof because it's their single resource because the carbon source is the same right but a question, a follow up question that I have is what happens if I increase the concentration of the resources or the daughter so I don't 15:37:50 have an answer, but it's something that I'm, I have to test because it's a it's a question that you want to ask, 15:38:00 follow up on these questions. What if I postulate that diversity in your system is proportional. 15:38:09 Sorry. The something else determines the diversity system. I'm going to postulate that it was the same in all of your tubes, and that the abundance is of individuals scale with the amount of resource you put in. 15:38:24 And now I'm going to post today that you have to shadow of a reading of sequencing depth to capture the abundant ones. So now that would I think my head. 15:38:37 That would explain why when you're now split the carbon between to resource you get the average abundance. And it will also explain why if you change your cut off your trends still remain the same. 15:38:47 So I guess it's basically the question of than if you were to when you combine two resources, instead of having to do 5050 you just to glucose the same amount as you did before, and the other resources mr. 15:39:02 monkey did before so twice the amount of carbon. 15:39:04 Would you then still get that average so I did a pilot study, trying to have the single resources and then having some so I had 60 resources and 16 plus plus plus the diversity was still higher, so I don't know that it was, again, higher than. 15:39:23 then single resources, but lower than there is the diverse that I had here. 15:39:31 Okay. And then shallow read, honestly the covenant was pretty good. 15:39:36 But again, this is relatively 20,000 on average, and the lowest ones how many restitution thousand. 15:39:48 I excluded those that were like 2000. And I had the lowest from 5000 to know I meant the lowest abundant or to use. 15:39:57 Did they tend to be represented by either. 15:40:03 Alright, thanks. 15:40:05 Tina what one more question for maybe tell me if you will get to this later. 15:40:11 I have a mental model of what what is happening here which of course could be wrong but to know if this is wrong or, or not, one will have to know what's the composition of the system so in my mind let's say there are a bunch of cross fingers, this is 15:40:25 you have 20 species, 19 or crispy there's one is using the substrate. If you ever noticed after you have 21, because they have another utilizing software but the CrossFit doesn't care anymore. 15:40:35 Let's get this I mean it wouldn't be a slope of 1.4 but we what we want and, okay, not so different. Is this what you think is happening. 15:40:42 I think, I think that what what's happening is a mixture of. 15:40:47 Yeah. What the species that are present, and metabolites is probably are secreted but at a low concentration that cannot be used the new stuff so it's a mix of the fact that you don't increase the concentration of the resources so basically at a certain 15:41:06 point some things go below the threshold and it can and cannot be used. 15:41:11 But this is something that we have to test, in the sense that I, I set up an experiment to test whether okay if I increase the concentration as Danny was saying, what happens. 15:41:24 And my simplistic model would be easy to check with the composition because it will be a block of things are concerned yeah now the composition was not really informative in this sense. 15:41:38 Okay. 15:41:40 More questions before I move on. 15:41:44 Ok. 15:41:47 So, again, we couldn't really understand the standard, we came up with multiple explanation no one was really working, or at least if you try a bit of modeling, usually get it. 15:41:58 It's a plateau in the, in the richness, with the number. If you are the more resources. So we went back to single resources and try it again to understand better this idea of course feeling good, and, and how the in in the parallelism between the metabolites 15:42:16 that are possibly produce the and how the species that are distributed. 15:42:19 So, this is just a theory that I'm talking about the metabolites that we estimated through the scope expansion analysis. 15:42:26 And what we did was trying to. 15:42:29 We look at how metabolites are distributed across the resources so. So basically we asked Okay, how many metabolites for intermediate. 15:42:50 And basically, based on these math. The common myth out there are some metabolites that are commonly produced that came from the image or the resources that you provided in the media in this for us are the immediate service you see a cycle in the lower 15:43:04 And then there are mythical ISIS instead are produced starting by just one or two resources. So this is a very nice by Mondo distribution. And you can see that from our prediction disagree quality economic resources can support, it can only basically 15:43:11 growth policies acetate. 15:43:25 provide the core metabolites. And we did the same thing for species and so this is the distribution of this, of all these, these that we found the single resources. 15:43:34 And we asked okay what is the resource of coupons to have diseases BC. And so we could find that some species factory like metabolites and are present in all the resources they know the community supported by other sources. 15:43:49 They know the community supported by other sources. While, and we call them arbiter generally some, while other species were instead associated with just one, or fewer source. 15:43:58 And also there are some intermediate valves that instead were found between Theon Tarbosaurus. 15:44:04 And when you look at the distribution of journalist intermediate and specialists realize that well in these broken organic resources proportionally that are more generalist. 15:44:19 And the other thing that we did before starting to draw in conclusion from from that was to say okay, have these specialists in general is different, based on some characteristics and the first well generalist mostly belonged to species to families that 15:44:35 now here you don't see the families but trust me. They belong basically to upsell them on adoption and enter bacteria. Bacteria. 15:44:44 And these are also, there were also most abundant there. 15:44:49 And the other thing we started looking Okay, can we identify whether the generalist have higher growth rates, or maybe there are more metabolically flexible compared to a specialist. 15:45:01 So what we did them was first looking trying to infer growth rates based on tabulated data. So we went to into the otter and DB database that provides the RNA offerings on coffee number, which is, which has been found that correlates positivity with the 15:45:22 maximum growth rate of bacteria, and we basically computed. We brought the irony coffee number of the species that we had this thing go to sources, and to find that on our very generous had to hire copy number compared to specialist, which can be indicative 15:45:41 indicative to the fact of the fact that journalists can tend to have higher growth rates. 15:45:47 And the other thing that we did was to try to reconstruct the genome from the sequencing sequences and try to see whether the generalist had the larger number or metabolic cancer. 15:46:06 And what we found is that this is true. 15:46:09 So basically, our species, the species that we identify this arbiter journalist and specialist differ based on some characteristics. If you don't stratify the data, have been the data this way. 15:46:21 Very, I just trying to understand your bins because in principle you could have just because you have three to 12 you could have just made those graphs for know I tried a bunch of different meaning, it's kind of the same. 15:46:35 Okay, doesn't really change that much. Okay, I didn't, I didn't do what like one than the others but basically kind of the same beginning is not in the beginning here is that one resource, and then it's just deciding where do you set the cut off between 15:46:51 In the beginning here is that one resource, and then it's just deciding where do you set the cutoff between the generalist and specialist. 15:46:54 No no I just. 15:47:03 Oh, yes. So, just to try this is a. Okay. And another thing that we did them. Okay now we started this generalist in specialists, and we are here, assuming that the specialists have a strong relationship with their resource compared to the specialist. 15:47:18 And so we asked the can we can we test the weather, one species is particularly loyal to is the resource based resource. And so we co created what we call a resource specificity score, which is basically a way to measure the tendency of species to be 15:47:37 associated with the particular resource. 15:47:41 And this is so you basically calculate the probability of a species to be in the old movies or the multi resource environments in which these species in the that contain the resource in which the species was found and interoperability. 15:48:00 This, the species is present in all the multi resources environment that didn't contain the resource in which the species was originally found you make the ratio of these two of the for the be one minus p divided by one plus B to C The second conditional 15:48:19 probabilities, and you get this ratio that is one when the when the species is always present is minus one when the species always obscenities target resource, any zero when it has no specific. 15:48:38 So basically when the species is everywhere. And what we found is that this is for multi resource environments that the generalist are present physically everywhere so they don't have a specific association with their particular resource wise specialist 15:48:54 instead. 15:48:58 And have a stronger, stronger specificity for the resource. 15:49:03 Is this clear or was it to foster. Okay. 15:49:08 And the other thing is, what when you look at the distribution of Johnston specialist in most resource environments. 15:49:15 We found again that there is the score number of species that is always present. And then the diversity did increase in diversity is actually driven by the intermediates and specialist. 15:49:43 I think you 15:49:43 know not always been present, but always been associated present when a specific resources present. 15:49:57 Yes, 15:50:06 wouldn't be here if there was a negative number would mean that species avoids the resource, so it would be for instance, if this resource kills the species, then you would expect to see a negative number, and you have almost no such cases maybe one point. 15:50:23 Yeah, right. 15:50:30 How many specialists you have on average per substrate. 15:50:36 That's Yeah. It is the plot is going before. 15:50:44 So, this is 15:50:44 no no this is the actually so the deal, bars, is the number of specialists Barry source. 15:50:56 Yes, he's like one that. 15:50:58 Yeah, so the total number of species that we saw in single resources was wonder than 70. 15:51:06 But I mean from that you have to subtract the generalist right because these are the crispy there's that. But they generally start these the pink guys, but then if you if you if these specialists are really all for example, responding to a particular 15:51:18 sugar. Then, now we have an interesting problem where we cannot explain that diversity, with breastfeeding. 15:51:25 I mean maybe I'm making too many steps in my head but they are the ones that are personally consuming the sugar. 15:51:34 Yes, yes agree. Like, 15:51:41 I don't, I don't, maybe I don't know if I thought 15:51:52 Did you try to make a scatter plot of the blue bars on the left, and light blue bars on the right. 15:52:01 What I am maybe I'm kind of seeing things which I, which are just spirits but for instance the resource which is, or they're not in the same order, right. 15:52:11 Oh, I see. Yeah, there's definitely that I definitely imagining things. 15:52:19 Okay. 15:52:22 This is separate Can I ask a quick question. 15:52:25 Yes. 15:52:28 Sorry my first zoom question I'm adapting a. 15:52:33 When I look at the richness in the single carbon sources which you have in the lower right corner of the slide, I noticed that tapped out at about 40, and you get about the same richness in 15 carbon sources. 15:52:45 Yes. So I'm sort of curious. In what sense, the multi carbon source communities differ from the single carbon source communities. Is it that a single carbon source is actually dominating the assembly of the multi carbon source community. 15:53:00 No. So actually, I can show you this later because I don't have in the presentation but as you increase the number of resources they commit become more even. 15:53:11 So, basically, you can have in single resources some tasks that dominate this is mostly for the glucose organic degree politic much less. 15:53:22 But then when you go there, what really distinguishes the single and the 15 is not is not really the richness, but actually the evenness. 15:53:35 There's fun to your question. 15:53:38 Yeah, thank you. Thank you. 15:53:42 I'm okay. You can try to move on. 15:53:47 So, I was here. We this. Okay, so basically what what what do you make of this picture of results which can be what for this for us we're not super clear immediately because here on one side you can think that while you have generally set are everywhere 15:54:05 and specialists that instead are just a few resources we can imagine journalist, being like being able to come to consume everything. But just be the general is scroll faster, can be. 15:54:19 But then, how can be. 15:54:22 The, the other species survive on the other resources, then the generalist grow faster, that varies on which resource still how fast they grow but there still will tend to be always there. 15:54:35 Yeah, but what they do consume. So the question for me was 15:54:40 the question that I had for the day. It was relevant for me was okay, other journalists consuming, should they think about the journalists as good as consuming everything, or should I think about the journalist as a specialist for a metabolite that is 15:54:55 always present in this case can be for example if you see an intermediate because they in our it yeah there are always produce starting from the resources that we provide income. 15:55:05 So, and we after we in then decided that this might be the way to think about this problem, and we put these in a consumer source model. 15:55:16 Okay, so the consumer source model between implemented is the classical MacArthur so fun cash style but the, the substantial difference here is that instead of having it around them matrix of breastfeeding. 15:55:33 We use the these maka with the carrot metabolites to inform the recipient verification question so that is the definition of specialists that it can't grow on the other resources or that if you put it in this community. 15:55:48 It doesn't survive it gets out competed for me in this particular context, especially, is it species that is found, just in a resource or some of those specially species and put them in, just by themselves in the various resources, do you know which ones 15:56:04 those will grow, I don't know. 15:56:23 I think if you if you also if you decrease the solution or change the illusion amount you can also sort of go from you know the the dilution is definitely important. 15:56:37 Okay. So, um, I was saying that the, this is a classical consumer resource model, but instead of having a random matrix of proceeding. We use the Kismet, the, the mapper that we derived from Canada to inform the crossbreeding matrix. 15:56:56 And, as I told you, a few minutes ago we decided that for us, or the box that we had in our community. We're specialists, either for different metabolites to those that are produced readily or four core metabolites, and these distinguish between generalist 15:57:12 and specialists. 15:57:22 This is a slightly different definition right. 15:57:25 Even as an isolated can only grow on. 15:57:29 Yes, this, this is in the model, everyone grows on a specific thing that can be either surprised resource or a metabolite to experiment you have in mind and sometimes behind this definition is growing isolates on the sources. 15:57:42 Yeah, the idea is that the certain point actually not a certain point I have the isolates now and I'm going to try a bunch of things, but for for this project I didn't have the isolates. 15:57:55 And so, I cannot tell you exactly how the real definition of specialist, is was I don't know 15:58:07 how many of them can grow in the shallow cities I can make a guess very few. 15:58:18 I can show you something later about some correlations that we see in this data between particular families and particular resources and their concentration, but this is just what they can infer from this data. 15:58:20 Yeah. 15:58:33 So, we're okay. 15:58:36 So, the way we constructed this matrix is basically saying okay. Each Bob is a specialist and consumer specific resources, a specific resource that can be provided or prospect them. 15:58:49 And in this case, so for example I have these bugs that eating. 15:58:54 So close them. And the way democracy is constructed is that the the bug is leaking just the intermediate center, immediately addition to the consumer resource in the metabolic pathway. 15:59:08 And then there will be other bands to specialize in these liquid byproducts, and they will release, other intermediates again downstream on the pathway. 15:59:22 and in so on and so forth. And so, by, and what else okay. The other thing that we implemented in this model is that we had a sort of, we tried to implement them as the idea of 15:59:36 that some metabolites are always produce the for example acetate site so we are this fixed amount of these metabolites to, to the model. And basically, once we introduce these structure in the speedy mathematics, I mean the simulations we got this linear 15:59:56 trend. So basically in the simulation we could see the, the richness the instinct. The richness could increase linearly with the number of supplied resources. 16:00:06 And another thing that was good do was OK, let's say that now with the species that survive at the end with India calibrated communities, I can again classify them based on where, in which resources they can be found. 16:00:36 this to defenders to the model, you have. So if you have a chain, and metabolic network and one uses the first substrate it leaks the second, and the second YouTube the utility of the second software would lick the third one yeah and so forth so you never, 16:01:01 The only the next one. Yeah. The next one or the multiple, next one. It doesn't mean to be a chain doesn't have to be a chain, necessarily. 16:01:07 Yes. 16:01:11 Yeah, so there's actually the throwing was quite accurate in the sense that you can have benching and then you can be. 16:01:28 Okay, so maybe one thing that why we can see the linear increase when actually they're not the number of metabolites basically can increase is no linearly with the number of resources. 16:01:35 And the way. Yes. 16:01:37 Sorry to interrupt but is that critical for the result. What is this thing that yes you always look because it sounds realistic to me. Yes it is. 16:01:47 It's kind of unrealistic in the sense that I agree, but the idea is that it was hard to simulate this idea that basically there is, is a community of specialists, and we in the general is that we have are just everywhere. 16:02:03 I don't know if I can do this on realistic to me was that along the chain you will with like the next thing because we know i mean that's not a word you have to run some part of the whole path. 16:02:18 I understand that is not completely realistic. And I agree with you, but that you've read something. 16:02:24 Perfect. I'm not, I'm. 16:02:27 I'm. Yes. Okay. 16:02:29 Yes. 16:02:31 This sounds a lot like this model toy model which actually that I consider three years ago. 16:02:37 Question is, video taken take on that, if you really are trying to build a multi species community out of it that the lower down this branch and network you go the smaller hours of Fluxus. 16:02:51 Yes, and in our model that was a major limiting factor for diversity at least in the key stats that you cannot really go more than two or three steps down because you're fluxes become just too low to support. 16:03:06 Actually, if I can comment on this basically the reason why we see the linear increase instead of the, of the plateau is because of the certain since the concentration of each method but I think this is one of it. 16:03:19 Then basically saw some metabolites. 16:03:32 Just to follow up another question. 16:03:34 So you tried the same thing but without that network structure and it does and you don't get these results now. 16:03:44 This 16:03:44 goes to understand. I don't even, I don't. 16:03:47 Okay, I'm not 100% sure that what's going on is what we depicted in this model, but the only thing to fear should is the observation that we have. 16:03:57 Okay, so 16:04:01 now is, I wasn't unsure whether it's one this, but since a lot of people here thinking about course screening. I thought that maybe this would be interesting to discuss we haven't really looked into this a lot. 16:04:15 But one thing that I want to show you is that even if you consider the number of families, you just found them as a function of the number of resources you still get in here. 16:04:25 And the other thing that is related to concentration maybe can help respond to repeat the question. 16:04:31 Daniel hand. 16:04:34 If you look, so we were able. So, if we found associations between as V's and resources, but also we also found associations between families and resources like other papers. 16:04:47 Other studies have found. 16:04:49 But the thing is that we were so we identify the associations using a machine learning algorithm, but this is not really important. The important thing is that wants to block the, the families and the abundance of this family is a function of the concentration 16:05:05 of the resources in the, in the environment. 16:05:09 And you plot, just those that are identified by the algorithm, you find that that is a very nice, almost think almost incredible to see how some families follow increase. 16:05:22 So the abandoned the relative abundance of our family, increased with some specific resources. 16:05:40 So for and just to give you an idea. These still don't want adoption and through what the ashes have all these January so the the species that we identify this generalist mostly belong to them and that's an introductory Archer. 16:05:44 And so you can see that they're mostly associated with the hydroxyl humor as well. And when this is reminiscent of some of the things that have been found by Alberto, on the other screen that these these the contrast between so the one a documentary about 16:06:03 So basically, for example, in, where you have a lot of demand action you that you have very few interrupted reaction. And the other thing is that for these families mostly found in Santa Rosa, and he's almost present present, only when you have sellers 16:06:19 sellers as a single resource. 16:06:21 We haven't looked into that much but I thought that this might be interesting, or at least a good point of discussion, if you have ideas about that. 16:06:31 Yeah, so. 16:06:36 Okay. 16:06:39 Yeah, I don't. 16:06:57 but actually the way we should think about it is that the Crispin network that is seated by this resource is actually responsible for all the changes that we see in the diversity of the communities. 16:07:11 And that when you think that it might be a. 16:07:21 Right. This was just an attempt to introduce the corresponding network in a, in a consumer source model so we're away from the democracies, and try to include the some of these ideas of metabolism. 16:07:32 And this was actually a game changer for us in terms of how, whether we were able or not to reproduce our experiments that result. 16:07:41 And this is actually that's actually makes me bit off. Make me a bit awful about the fact that we can actually use resources there tapping number to control an engineer communities, busy and obtain states that we desire and. 16:08:03 Am I doing with time. 16:08:05 Okay, so maybe I want to talk about some open questions that we may end after this project, there are a lot. 16:08:15 And there are a lot of things that you understand to be fully honest. But one thing that, to me, was extremely important and extremely interesting is that one way to think about resources, is to distinguish between great quality can look on your journey. 16:08:36 And I want to show you this I haven't, I haven't included before, but basically here. 16:08:53 Before I was showing you how the richness varies with the number of resources. 16:08:54 You can actually distinguish between the number of broken organic angle equality. In this case we had just three click blueprints and resources but still you can see that think this is diversities you only to the electronic resources. 16:09:04 Because when you are the group interchange resources, diversity doesn't really increase. 16:09:10 And I worry that this makes sense because basically the cyclical brokerage and resources can only provide this quarter metabolize, and only this generalist can really survive, excluding the specialist. 16:09:23 And obviously this is in any part of this that can be tested experimentally so basically can repeat the experiment, including more broken organic resources and see whether this is the case or not. 16:09:37 I'm still a little stuck on the, on the previous bit. 16:09:41 So, so this this algorithm that you outlined for figuring for making the consumer research model. I think this step that you do where you you say what is the next thing that gets made it makes a lot of intuitive sense to me for things that are metabolized 16:09:52 extracellular early. 16:09:54 So like for cellulose or something like Yeah. 16:10:08 something of that nature. And so I'm, I guess I'm really curious what the what the sort of control model was like what was the other thing that you did that convinced you that this was the way the random matrix. 16:10:21 It works. 16:10:22 No, I didn't. 16:10:25 Okay. 16:11:16 Are you sorry I interrupted you. So the comparison was the random matrix thing was there another yeah we started also when you leave everything. And this was not really working. 16:11:23 And I think so me. 16:11:26 Yes yes okay was it. 16:11:32 Yeah. 16:11:35 Also I just wanted to follow up on this, and when you say worked or didn't work. You did not make an attempt to feed those parameters to the species you observe he just says that if you make those strong man species which are specialists in one resource 16:11:51 and exclude the next one, you get a similar trends of diversity with, with the number of resources. 16:12:01 Have you actually tried to be a little bit more realistic here, again, what what you did is great but maybe when you are thinking about the next step. 16:12:16 If you try to argue that here my species. I have not just presence or absence actually have the steady state abundance data. 16:12:21 Can I try to make consumer resource model. 16:12:34 Include in some unknown secretive metabolites, which would be consistent or best correlated visit data so that would be something I would try it next. 16:12:50 Okay. 16:12:50 Well, it's a batch dilutions it's it's it's not a big difference. It's not a huge difference you can if you, if you have your parameters of consumer resource model a maximum growth rates. 16:13:02 you can almost have a precise mapping between the two. 16:13:05 So I don't think it's a big deal. 16:13:11 Ok. 16:13:16 Okay, so the other thing that I think it's interesting to think about them is that these communities that are going on different resources are different for many reasons so they have different weaknesses, they have different even as, but also you can 16:13:32 you can look at other feature for example the average size, whether these metabolic overlap between the different microbes that are inside the community, and one question that I have and I think it's interesting if it is from my standpoint is that whether 16:13:48 this feature can be predictive of the community responses to perturbation. 16:13:55 And to give you an idea of what I'm telling you. So here, for the single resource communities that we could calculate the average user is the endemic metabolic overlap. 16:14:07 And you can see this we have a range of gym sizes, the average minimum size of the box in the community, the metabolic overlap the community and this correlates negatively with the richness, but actually this I think is a good platform to ask whether 16:14:23 communities are different for these reasons, maybe respond differently to different motivations. 16:14:33 And finally, the other thing that you already mentioned that I think it's crucial is the concentration so previous work from the lab has shown that when you increase the concentration, you actually increase the, the frequency, the strength of negative 16:14:46 interactions. And so your speculation is that you might decrease the richness of communities, but basically we don't know what happens when you provided the concentration. 16:14:57 And what happens to our. 16:14:59 So this is a completely open question, and says, These are the avid these averaging this we get from Steve from going from one to two resources, kind of a big, it is not that we cannot really understand maybe taking into account the concentration, and 16:15:16 destiny experimentally can provide the answer. What was the nice thing about all these question is that they can be tested experimentally. And so that's what I'm doing now actually have some data about this, that is not in the form this could be presented 16:15:30 but I would be happy to discuss them with you if you want to receive your feedback. 16:15:37 Yes, so you just question how do you measure metabolic overlap. So yeah, okay. So basically, again you can reconstruct the genome, through pie crust. 16:15:48 And then you can, and then you can calculate the similarity, which tells you whether metabolic genes are presence or absence and just say, Okay, how much, how much overlap there is between the box of this community compared to the others based on the 16:16:03 metabolic genes that are present. These are reconstruction so 16:16:11 a metabolic you met you, you focus on metabolic jeans. 16:16:16 you bet you you focus on metabolic jeans. The thing is that these are reconstructions, and I, I think I will do the planning to do some shotgun, to have the gene. 16:16:28 Do that, as he said we do have a slide with abundance data, just because I was thinking about whether you can say something meaningful by meditating on those abundance distributions. 16:16:42 Do you want to see abundance for some of the resources. 16:16:47 Anything which kind of is representative, not the exotic cases and 1000 different language between primary concern, consumers of primary resources and secondary sources from the abundance. 16:17:04 Presumably is a secondary resources will give you a lower dramatically lower octave. 16:17:09 I can tell you that for example these in these cited Fumarate enthralling, the highest abandoned at London's was of the. 16:17:19 If you have a slide that no I don't have a slide I can, but one thing is. 16:17:32 Okay, so it's not completely the full picture but basically if you take the relative abundance of the soda Manas in the Roxy broad enough numerator, this is the most abundant species. 16:17:42 And then the others are just very, very tiny while if you get cellulose for example, you see that, for example, these communities, the salvage international which is actually one species. 16:17:57 It's a fairy but it's represented by one species. It's not a barrier, it's, it's there, it's not very abundant but there are a lot more species, and we more even abundance. 16:18:11 The insoles good that's actually, I was thinking about some time that exists, you can assign them a specialist for each of the resource for instance. Yeah, you know that for humor rate this brown circles species, whatever it is, is really the primary 16:18:31 consumer everything else is just based barely scraping by, and on it's not the way to read this blog, though. 16:18:34 Oh yeah. Yeah, it's a family. Okay. Did the circle is just the abundance. 16:18:41 When there is only one, this is when they're bonded. This is the abundance when you don't have it. 16:18:48 And this is the seven degrees the abundance when you have either actually foreign together with the many other resources this the abundance of the species when I Roxy crawling is the only resource that is provided. 16:19:01 Right. This is the way to interpret yeah but I was thinking about the humor rate, and then then it's as as Pankaj pointed out, this is a family of, I don't know how many. 16:19:14 But, yeah, but I would say that this this as another species but I would say that there is basically one or twist, his visa, that belonged to the monitor that without a drive in his abundance. 16:19:28 So as you go along the x axis, does the concentration of hydroxide proline change, or you're just adding more resources or the more resources. Okay, so the content Okay, yeah, that's. 16:19:46 All right, we're going to band questions, including mine, so you can tell us whatever you want to tell us or less you don't want to say anything about temperature for 10 minutes, I can, I can keep the passion, it's fine. 16:19:53 We can keep on discussing. Okay, if you want, if you prefer 16:20:02 a question or suggestion I guess for for of experiments. 16:20:07 Are you able to add 16:20:13 the in the hypothesized or suspected intermediate sort of downstream products to YouTube's. So not just the upstream initial carbon sources but anything you suspect downstream might also be acting to feeding some. 16:20:29 Yep. Yeah, seeing whether indeed. And how many of them are actually limiting the growth of at least one buck and or whether any of them limit more than one buck, so I wonder if in ultimately if you take everything into account. 16:20:49 If you can, again, verify if it actually is consistent with the competitive exclusion principle after all. Okay. You have to actually check how many limiting resources are actually there and limiting means increasing by little bit, somebody goes up. 16:21:14 I guess the difference is whether you let them saturate, right, because if you wait for them to saturate. 16:21:21 That doesn't matter. That's true. Like, if you have a 24 hour they reach saturation, right, 24 hours they reach saturation, so they do reach saturation Yeah. 16:21:44 I'm just trying to understand the point. You two are making. I think the point is that if they reach stationary face, there's a lot of things that leak out, because it says are starting to start and weird things happen and then they leak out everything 16:21:56 so it's more and model where the leak out is very. 16:22:00 So this is like a terror effect. Oh, are you you're worried about terror effect okay God I got. 16:22:10 There's a time dimension, you can get more things coexisting. 16:22:20 Leave the really leak out everything which is your. 16:22:31 And how much. 16:22:39 That's another field that's big enough. 16:22:47 That's a big enough phenotype space and there's no intrinsic elements. Now whether under some reasonable model and evolving and so on, you can actually get a lot of species there is not clear but it's just, it's intrinsically different the growing in 16:22:52 That's different than if everything's there and steady state, it's just it's so I don't think that the, and I'm, you know, even this this division into the two classes behaving differently that it may also be to do with overall girlfriends do under those 16:23:15 resources are some of the generalists, so I'm not sure I think somehow has to put more of a summer old gn to explain what's going on. 16:23:28 So I put a competitive exclusion limit just the same way as humans that would be bad take another question. 16:23:41 So, so let's forget for the moment about Daniel's objection. So in a consumer resource models, you know I don't work in them but the one sort of result I know is if you have, you know, 10 resources you at the end get us most 10 species like coexisting. 16:23:55 But this is kind of a question for the theorists. Is there some idea on what among the species that coexist, what they're sort of distribution of relative abundances. 16:24:06 Because that's sort of another way to try and compare to this data. and it also kind of relates to. 16:24:11 If we expect things to be appearing that different relative abundance by five orders of magnitude than these measures of richness are fundamentally flawed then to because we're not going to catch those very low abundance things. 16:24:21 So So do we have an idea on what like the distribution of relative abundance is is and consumer resource models.