 Good day. I'm Max Hegblom, Editor-in-Chief of FEMS Microbiology Ecology, and it's my pleasure to welcome you to this webinar on microbial life strategies. We continue our webinar series, which enables us to highlight different topics of microbial ecology to a broad audience. Those of you who have tuned into our previous webinars knows that FEMS, the Federation of European Microbiological Societies, invests in science, using the income from our journals to fund charitable activities and support our community, such as providing grants to scientists, organize and support conferences and summer schools, and sponsor a range of events such as this webinar. And this provides a forum for presentation and discussion of key research enabling the flow of ideas to a worldwide audience. So by publishing in and reading FEMS journals, you support these activities. If you missed earlier webinars, they are available on the FEMS OUP websites. Indeed, I wish to thank the staff of FEMS and Oxford University Press for all their work behind the scenes in making these webinars happen. So today, we focus on microbial life strategies. Microbial communities vary greatly as microbes adopt to different niches. Functional and physiological traits are underlying drivers of this niche differentiation. Common framework is based on copiotrophy versus oligotrophy, where resource investment is primarily in either rapid growth or stress tolerance respectively. Predatory protests are also major consumers of soil microorganisms and the feeding patterns by these predators have an important role to play in shaping soil microbiota. It is also recognized that microbial communities can have dramatically different compositions, even among similar environments. And this multi stability and multi species communities combined with ecological noise may lead to unpredictable community assembly, even in very simple environments. Indeed, there's much to learn about the factors that shaped the composition and activities of microbial communities. Our three speakers today will dive into these complex topics of microbial life strategies. First, Damien Finn from the Thunan Institute of Biodiversity in Germany will discuss functional trait relationships. Then, Natalie Amaker now at the Swiss Quality Testing Services will examine protest feeding patterns. And then finally, Eric Wright from the University of Pittsburgh will discuss multi stable bacterial communities. After the talk, we will open the session for questions and discussions. So please submit your questions via the chat link or the Q&A link and we will get back to these then at the end of the session. So with that, welcome everyone. And first of all, Damien, I'll give you the floor. Thank you very much for that. Wonderful introduction Max. Okay, hopefully everyone can see my screen I have a little message saying that that's so today I would like to discuss oxotrophy as a life strategy with you. I don't know if I'm using the term life strategy a little bit too liberally here, but we'll see how we go I'll let you be the judge of that. So when we often think about evolution, at least me personally, I tend to think about like, you know, the new function coming along or at least the adaptation of a pre existing function to new set of conditions. But that's not the whole story. And gene loss is a pretty powerful driver of evolution as well. And generally, if there's no selection pressure on a gene or a function that it provides, then mutations can accrue over time with the final result being the loss of that function. And so we know from the long term evolution experiment that if you accept a successively grow E coli in a medium where glucose is the sole carbon source, then over time those E coli strains will lose the capacity to catabolize other carbon substrates because they haven't encountered it before they don't need it. It's just it's not important anymore. Also, if we go out into the the actual world into the wide blue ocean here, we find that some of the most abundant marine taxa floating about in the surface layer of the ocean also are riddled with gene loss. A really good example of that is the loss of catalyze in pro Thoracoccus. So the tricky thing is, when gene loss sort of means that you've lost some kind of protein or growth factor that you require to live. And what we do know is that, yes, absolutely, the ability to synthesize growth factors is susceptible to gene loss. We call prototrophs organisms that are able to synthesize all of the growth factors that they require. And then conversely oxotrophs are those organisms that cannot synthesize all of their growth factors. And just like in our E coli that only grows with glucose over time. If we grow E coli successively in media where all the growth factors they require are provided to them, then they will develop oxotrophy. Again, you walk out into the wild blue yonder out into the world and we see that oxotrophy is common across prokaryotes, eukaryotes. So you are an oxotroph. If you decide to take a long sea voyage from Europe to Australia without packing any vitamin C, you will get scurvy. Incidentally, this is by far the least nauseating picture of scurvy I could find through Google. I would not recommend it based on those pictures. But also in prokaryotes with some nice work done very recently, we see that in environmental culture collections, upwards of 50% of those strains are oxotrophs as well. So this loss of growth factor synthesis is out there. It's a thing. Yeah, so the problem here really arises that once you're as a taxon, you're locked into being an oxotroph. You really need to think about how you're going to get the growth factors that you require. Otherwise, you will die. So a popular choice for a lot of oxotrophs seems to be to become a parasite or a pathogen. And so by that way, you can forcibly take the growth factors that you require from a eukaryotic coast. Alternatively, you can be like the friendly, obuscular mycorrhizal fungi and you can become a beneficial symbiont. And whereby you can do a nice job for a plant by picking up phosphorus from the environment, for example, while the plant will sort of give you sugar in return. Both of these examples are microbes with a eukaryotic host there. And what we know as microbial ecologists is that the vast majority of biodiversity meant most of the taxes that actually exist are all free living out in the oceans and the soils, etc, etc. So the question really arises in these complex communities of free living organisms, who are the oxotrophs and how are they getting their vitamin C through a various collection of my own poor life choices. So I work in the soil environment, which is a little bit tricky compared to some others. And so all of the work that I show you today, there are some things that we have to keep in mind. The first of which is that soil is a very heterogeneous environment and there's an issue of spatial scale there. So there's separation of not only microbial taxa from each other, but there's also separation in terms of the resources that they require, whether that be water or gas matter. And so you may have read or heard the sort of the very lovely phrase that like, you know, a gram of soil is a universe, there's almost a billion cells there. That's very true. However, the vast majority of those taxa will never see each other, they will never interact. So in the soil environment, already we're starting to see that it could be a very, very risky sort of strategy to be an oxotra. Furthermore, all the work that I talk about today, we work on the other scale of a soil aggregate, which are these tiny little sort of like three milligram amounts of soil, about two millimeters in diameter. And whilst that spatial scale is still a lot bigger than what is realistic in terms of microbial interactions, it's as best as we can do and it's better than working with 250 milligrams, at least for our questions here today. Another thing to keep in mind is that there are different micro habitats that exist in soil. So there's the bulk soil, which is a relative desert in terms of the amount of carbon input material that comes into the system. And then you also have the rhizosphere, which would be much nicer for microbes because on the surface or immediately around the plant roots, they're receiving a cocktail of nice carbon substrates and also growth factors all the time. And so here already we can see that oxotra would preferentially want to be in such an environment rather than the bulk soil. So for the work I show you today, we asked these three specific hypotheses. So firstly, we hypothesized that there would be more oxotra present in the rhizosphere than the bulk soil. Secondly, that the complexity of an amino acid will be linked to the gene loss. So if you require more genes to be able to produce a certain amino acid, then just from probability, you're more likely to have some sort of mutations in there and the loss of it over time. And then finally, we predicted that for those oxotra that do exist, they have to make some friends, otherwise they couldn't exist there in the first place. So we took a lovely lime soil from Hildesheim in Deutschland here and some ZMA seeds from the inbred lion B73 and then with a very simple pot experiment. We had a bulk soil control with no plant. We had some maize plants growing happily. All everything was kept at the same moisture 20% for about two months. And then the maize was harvested at what is called BBCH growth stage 19, which is just after the establishment phase of the root rhizosphere of the maize and just before the maize hits its exponential growth stage. We took individual soil aggregates either from the bulk soil or from the surface of the maize primary or lateral roots. And these aggregates were incubated on a nice cellulose agar medium for up to six days. And so it's really great. The system that we have here, because the aggregates are exposed to these agar, we can manipulate the communities there in sort of interesting fun ways. And the way that we went about that for this study was to either provide those aggregates with the nice L form of certain amino acids that increase in complexity, lysine, phenylalanine and tryptophan or the toxic anti metabolites, which are structural analogs of those amino acids. So if you're unfamiliar with what an anti metabolite is, historically they've been used as a way to actively kill oxytruss and select them out of populations. And basically the micro thinks that the anti metabolite of lysine is the same as lysine. It tries to incorporate that into protein synthesis, but the protein doesn't work the cell dies. The treatments that we used here was a negative control where no amino acids were provided a positive control where we gave the good forms of the L amino acids of lysine phenylalanine tryptophan to everyone. And then we have three different anti metabolite treatments where one of the anti metabolites was substituted. And then afterwards we did a very typical microbial ecology workflow, DNA extraction, PCR and QPCR of the 16S RNA gene, amplicon sequencing of that, processing with Chin2. But we also did a shotgun metagenomics of the enriched communities and did that on the NovaSeat platform, and then those metagenomes were run through MetaRap. Okay, some results. So you'll be quite surprised to hear that when we provide amino acids to microbial communities, we stimulate growth. And in this sort of yellow curve up here in these plots, one from the bulk soil from one from one of the maize plants, that was the positive control where all the amino acids were provided. And of course, this makes perfect sense. These extra amino acids, they can either be an extra carbon source directly for the microbe, or they can be funneled straight into our protein synthesis. Where things became interesting were the differences between the different anti metabolites. So what we saw was that the anti lysine response was very, very slight. There wasn't much of a loss of the population growth relative to the positive control there. However, when the communities were forced to synthesize the more complex phenylalanine or tryptophan, their growth or even the stability of their presence over time was sort of kept much more similar to the negative control. And remember in the negative control, these communities actually need to synthesize everything. So not just these particular amino acids, but vitamins and so on and so forth. And then the other major interesting sort of difference here was that with all of our rhizosphere communities, we found that they were very poor producers of tryptophan, and indeed they were dying out over time. And it looked like these communities that had adapted to the rhizosphere were not able really to sustain their tryptophan needs. Looking at the functional capacity of these microbes. So this is just a heat map and it's presence absence based. And the vast majority of taxa that we could identify, they all encoded for the capacity to synthesize their own lysine. And this suddenly made the QPCR data make a lot of sense so it really is not surprising that the anti lysine had such a little effect on killing the microbes and that's because they pretty much all make wise. However, it got more interesting with the phenylalanine and tryptophan. And what we found was that the capacity for this synthesis was far more patchy. So here we have phenylalanine in blue, and then we have tryptophan down here in orange. And this is where we were picking up the gene loss in these taxa. And so really it should be the phenylalanine and tryptophan where we're starting to see these ecological interactions between oxytrophs and providers. In order to predict these interactions we turned to network theory. What's something that's really fantastic about this particular study is that we can build temporal networks here which in my opinion makes them a lot more robust. And the basic idea of that is that with every single taxon we know how it grows over time. So down here we have say for example the growth of taxon one and taxon two and we can see how similar that is. So we start to build our Spearman covariance matrix of this growth of these different taxa over time. What we're really looking at is how similar are these organisms matching each other. And then of course what we do with networks, we take either the very strong positive connections or the very sort of strong negative connections. And then we build some hideous thing that looks like this where every node is a individual taxon. Every red line is the weighted measure of how negative their covariance is over time so they did not sort of grow together. They went away from each other and then in green we have a positive growth. And then these are what our networks kind of looked like as a trend. When excess resources or the anti-lizing were provided to communities, we were always getting these very sort of simple, very few nodes. They were all organisms like pseudomonas and bacillus so these sort of like stereotypical competitive copiatrophs and they just formed these little sort of clustered blobs. When we forced our communities to synthesize more complicated amino acids or all of the amino acids, you are getting far more interesting complex networks like here. And basically with all of these we're getting more taxa but also all of these nice sort of sparse interconnections between them as well. Looking at these, it's an absolute mess. It's very difficult to make any sense of this whatsoever. And so what I did from that point was to ask the computer very nicely to run through the pairwise connection of every single taxon in all of these complicated networks to try to identify who are the organisms that are always making friends with each other, who are the consistently recurring partners. And then from that you were able to build these much more tractable networks that made a lot more sense. So in terms of the consistently recurring partners what we found was that here at least in this subnetwork in the middle this larger one. In the central point we were getting two individuals methyl oceanobacter and an uncultured actinomycetia. And from our functional information, we could say that these two organisms were fully capable of making lysine phenylalanine and tryptophan. So that was quite interesting. However, all of the organisms attached to them, well all of them but these four at least, attached to them in the periphery from our metagenomics work we could say that these were all putative oxytorps. They were unable to make either phenylalanine or tryptophan and they were all connected to our two providers. So this was quite an interesting outcome because it really sort of gave us some sort of confidence that yes we can indeed see who are the potential providers and also the oxytorps in terms of these growth factor cross feeding interactions. It's not perfect. There are some individuals here where either the taxonomic resolution or the functional resolution wasn't perfect. I can tell you why Vichynamibacteralis was connected to our actinomycetia but still picking up these particular general are sort of really gave us confidence that we're on the right track for these analyses. So I'm afraid that I've probably gone over time but just very quickly. But in conclusion, we found that amino acid complexity is absolutely linked to the frequency of gene loss. A taxon in the free living taxon in the environment is more likely to be an oxytorps tryptophan than it is for lysine. We found that there are more tryptophan oxytorps in the rhizosphere than the bulk soil and so if you have become an oxytorps for tryptophan, your chances of survival are much better if you start to make friends with the plant rather than sort of risking living in the bulk soil. We identified some potential keystone providers that probably play quite important roles in these communities for oxytorps and they were methyl oceanobacter and then this uncultured actinomycetia that kept popping up. And just in general, we could conclude that the need for cross feeding interactions is likely quite prevalent in these free living communities and having a high biodiversity that we often see in microbial ecology is probably good insurance for oxytorps that are forced to live away from a potential new carrier first. Because the more individuals that are out there, the better your chances of finding at least someone who can make the vitamin C that you need to not get your scurvy. So to acknowledge, I would like to thank Christoph Tebi and also Mia, Christina and Graham, one from Tunin, others from Echols on Trades de Lyon, all of whom are forever sort of supporting me and all the crazy ideas that I come up with. I would like to thank Max and Ben's microbial ecology for inviting me and then the DFK for monies and things. Thank you very much. Thank you, Damien, very, very interesting. I already have a whole bunch of questions also to follow up on and again for everybody else will leave those to the Q&A session at the end. So again, please type in your question to read a Q&A link and we will get back to these then at the end. So no, thank you. And we move on to Natalie Amaker who's going to provide a different aspect on microbial life systems and what's driving communities from the bottom up now to the top down looking at protest feeding patterns and how they impact soil material community. So Natalie, welcome. Hi, thank you. Thank you, Max. It was very nice presentation. So I hope I give it up to this level. Let me try to share my screen. I think that should be good. Right. Can you see it? Yes, perfect. It's always the bits of the stressful moments. So yeah, great to be here and to be able to present. Thank you very much for the invitation and the introduction. So I want to discuss today, predatory protest, and most of the work I'm presenting today is related to the paper we published with earlier this year. And I always like to first start with what protest are because you may not all be very familiar with them. So parties are eukaryotes and they are eukaryotes that are not plants, not animals and not fonti. That's probably one of the most terrible definition in biology, because it's like, it's all what they are not some is always trying to give what they are so they have their eukaryotes and like eukaryotes they have a nucleus and member and bounce organelles. This is different from prokaryotes like the bacteria. And you can also see here, I will get the laser pointer. So I can show you here. And here you can see so that they come in various sizes. So from a few micrometers up to millimeters, they have a great variety of forms and morphologies. So you can have amoebas, flatelates or some other very nice amoeba that does have a shell so they protect themselves with a shell. And so many more so I want this was all of them but just for you to appreciate a bit their diversity. And here so you just have a small amoeba moving around just for you to see so how they are when they are moving with all this variety there so of course have a lot of different ecological roles so we can find some primary producers in the group of the produce decomposers, parasites and predators. And in the soils, it, it has been shown that most of them, the most abundant functional group of parties are actually predators. And here I just have two graphs, showing that from from some recent papers, they both were looking at the distribution of functional groups of parties. And that in the soil, the most abundant group was the one of the predators or consumer, how they term it here. And of course, is always interesting to know who are the most abundant one but we also want to understand why is it important and what they are doing in in the system. And especially in my in my research, we were looking at how this may be related to plant growth and how the parties maybe predatory parties may be important for blank growth. And this is so due to some previous literature that was showing that the presence of predatory parties can really beneficially influence plan growth. And this was explained through two main mechanism. And one of them was firstly described in the aquatic system, and this is about microbial the microbial loop. So when you have a predators in your system, you get more nutrients released. And this is really a like in certain nutshell. So of course, if you don't have predators, you have bacteria feeding on nutrients and they keep the nutrients in themselves as long as they are alive. But when you have a predator, the predators is making their life a bit shorter, and some nutrients get released a bit quicker in the system. So everything is a bit fast enough. But there is also a second mechanism. And this is a change in the bacterial community composition. And this can be due to, this is a two size story. So of course, you have some bacteria or some prey that are better off, they are better at defending themselves. So in presence of predators, they will have another advantage compared to other bacteria that are not so good at defending themselves. But we also have some parties that you prefer to feed on certain bacteria and not the other. So, due to this, this choices of either the bacteria or the predators, the produce, you will have a change in your community. And this is something we wanted to look at a bit further. And we had a couple of questions. So, thinking back about the diversity that we have in protest. We wanted to see if we can see food preferences and predatory impact as at the tax and specific level of parties so our specific parties isolates. Do they vary in their food preference. Can we also predict that maybe through some measurable traits that we can have for this parties. And as you do this food preference relate at all to predatory impact in a more complex system. So what we wanted to look at, and we had to select some protest. Of course, I show you that there are quite many. We couldn't take all of them. So we had to make some choices. And we decided to take eight isolates from our collection so we had quite a big collection of parties and we decided to take some parties that were closely related in terms of the fellow journey and so a bit more distance related. So here you have the eukaryotic supergroup, we took four of them from this group, and each time we took two that were closely related so we have to a cantamiba and to vanilla for instance. They also had different growth rates and they had different volumes. So in this way we try to look at different traits, and if these traits may have a so impact in the feeding preferences. So I'm just showing you. So again, a bit the same story is just for you to to see a bit how far related they are compared to each other. So this is just a figure representation of a phylogenetic tree of the eukaryotes. I also put in transparency here the plants, the land plants. We have the metasoba, the fungi, and all the rest are put together under the term of protest. And here you can see we took the four parties from this big group, then we had two parties from this group, and two other parties from this group. So just that you have a bit of a novel view. And how we then started our experiment so we wanted to look into feeding patterns as I said before, and for that we decided to do a sort of to use a quite artificial system. We had our eight parties, and we grew them on 20 different bacteria that we were we also had quite some characterization for this bacteria. And we wanted to monitor how well the parties would grow on this bacteria as a proxy to say this food is a good for good food source or a bad food source for this parties. And this was a quantification by microscopy. And this is so maybe I don't know how many of you have counted parties on a microscope, but this is a type of quite time consuming activity. So that is so explained that we had only only eight parties here. So, just so that you know, but they are amazing to see under microscope so do so if you have a chance. And here just to show you a bit sort of the results already from this far from the feeding patterns. So you have on the x axis the different bacterial species, and on the epsilon axis you have the different parties. How to read this heat map is to see that when you have a blue color, it means that this parties could grow pretty well on this bacteria. When you have something more orange yellow, the parties could not grow so well on this bacteria. And this is the first data so that we got, and we wanted to understand a bit further what's what it mean to characterize it a bit further. And to do so we, we met a couple of analysis. So we looked at the coefficient of variation. And this is the standard deviation over the mean and we did it per parties, and that was to try to understand how I call it how peak your parties is. And so to see how, if you had a bit of the dietary niche press to try to to understand if your parties really only feed on one specific bacteria and almost not at all on the other. Or if your parties is a bit more generally stand feeding a bit the same on all of them. So that's what the idea behind. And we also looked at the Euclidean distance between the parties to see if they had similar feeding patterns if they were feeding similarly well on the same parties or or not. And I mentioned as to when we selected our parties we had different traits that we looked at. And there was the phylogenetic distance growth rate and volume. And we looked at this if these traits were correlated to the feeding preferences we got. So that's for this part and I'm just showing you a bit of the main sort of results characterization we got from there. And we found that most of the parties we studied here these eight parties were rather generally generalists. They had relatively broad dietary niche that there were no extreme parties would only fit on one bacteria. And then we found looking at the Euclidean distance this feeding patterns and protest rates that most of the growth rate was potentially important in explaining this feeding patterns. And this is one part of it so this part one so feeding patterns, how we can characterize them. The second part, we wanted to look at the predatory impact of produce. And for that we took the same protest, and we added them in soil microcosm in this time in five replicates. And we looked at the change in bacterial community composition via sequencing of amplicon sequencing. And this work has been actually entirely done by one of my colleague, Dilega, who is the co-author of this paper so credit to her for putting all of this together that's quite some work. And from that again, we tried to look at the we characterized the predatory impact. And first here just was searching a way to present to you the data. So I went for this rainbow colors. And that's the alpha diversity of our communities in the different treatments so you have the control and the different treatments. I won't stay here on this too long but just, you can see that you have a dominance for instance of proteobacteria in our system, then you add bacteria or data and actinobacteria, then. And similarly from the order from the feeding patterns here, we try to look a bit further and to see how we can maybe characterize this impact. And here we decided to look into the nearest taxon index, and this value this analysis can can help you understanding if you're having a phylogenetic clustering or over dispersion in your data. So if you look into that you see if we had phylogenetic clustering especially because we this would suggest domination of deterministic processes. And we thought of predation as something that would be quite deterministic, especially if you have one parties quite specialized into feeding in like we would really prefer to feed on certain bacteria. That way you would really lose this bacteria in all your system, given you have this bacteria in all the systems. And then we also looked into with, we looked at the breakout is dissimilarity relative to the control is to indicate if we actually had an impact of operators. So we looked into changes in community composition, compared to the control. And so, compared to each order to see if two different parties would have a similar impact or dissimilar impact. And we also want to go a bit further because we found this big better diversity information, maybe a bit too broad to really see what's happening in our system. So we also decided to try an analysis and to look into an analysis to make differential expression. So here I'm just with the dissect package in in our for the one of you who are, who are familiar. And we can see here the luck to full changes of the audience in the treatments compared to the control. So yeah, I'm only showing to only presenting two of these graphs I have eight one for each of the treatment, but we won't go into details in that. I just wanted to present them because it does give a nice view of how potentially each parties is having sort of a fingerprint of a predator impact. But especially if you look at this, which you level you may be able to, to find interesting, interesting. But the use that are maybe always increased with one parties are always decreased with one parties and I think we can go a bit further if we go into this direction. And so, that's also difficult to make sense out of this data so I think it's a nice direction to go and let's see how far we can go we didn't do much more here than to see. So there are different we can definitely see that at the, at the level of the issues, each each of the parties is having this sort of this sort of fingerprint predatory impact so he just to summarize a bit and to come back to the NTI because I didn't get a result there yet. NTI this nearest taxon index, we couldn't really find any clear direction. So we had for all of our, our communities NTI that was higher than two, and it was similar for all our treatments, including the control. We don't know exactly but maybe just our system was sort of deterministic in the way we, we did our experiment so didn't really help us to, to understand our system, but still a very interesting approach, I would say. So we look at the brecurtis dissimilarity and the log two form changes as I say, here we could definitely say, each of our parties is having a significant and distinct change in the back to our community. So that was very interesting for us to see. And now we wanted to go further so we have this feeding patterns, this predatory impact that we tried to characterize in different ways. And we want to see if actually do they, do they match, do they relate to each other in some way or another. And for this, we had so two different main hypothesis. It's a bit of a lot of text but I couldn't find a way to make it shorter so bear with me. We had this first idea that a produce with relatively narrow dietary niche breath so parties who would be maybe more bit of a specialist. This is what I'm trying to illustrate here with this. This would be a flagellate and really likes this orange bacteria but not so much the other bacteria. And we would expect that when we have such a situation then when we go and look at the impact on the community we would have a bit of a stronger impact and a bit more. Yeah, clear impact there because then you have maybe all than the yellow bacteria that were similar to the orange one that just disappeared for your system. While on the other side when you have maybe a bit of a broader dietary niche breath. You have a produce with just feeling kind of the same on all of them and then you may not be able to detect a very strong change in your community compared to a community without predators. So I think that we tried to to correlate the bracket is dissimilarity relative to control and or the anti I to our coefficient of variation. And as a reminder so our coefficient of variation is proxy for the dietary niche breath and the anti and record is dissimilarity are representing or indicating or sorry or proxy of the predatory impact the magnitude of the predatory impact. And as you can see here there is no big relation. So, but that doesn't make sense, as we saw that offer this we're all more or less generalist, and we didn't find a very strong direction in their in their impact. So, in that way, I think this hypothesis we can say that it was very rather not what we were expecting. It was not exactly surprising I think so we can expect this to happen. But then the second one was to look into if we have a similar produce that have a similar feeding patterns that they do this part is to have a similar impact on the soil bacteria community, because that would be so very interesting in various ways I will say more about it in a bit. So here we did find that that was the case so we did find that produce with similar feeding patterns did have similar predatory impact. And this might not be the best figure to show. And if you have some suggestion I'm happy to, to get them I couldn't find a better way to present this stage I think. So this is a PCA using pairwise distances and here we are explaining the variation in our, in our communities with different trades different variables. And we can see that the feeding patterns the growth rates and the predatory impact are all clustering together to explain more than 50% of the variation in our system. And then we have so the volume and distance phylogenetic distance explains explaining on another axis of our system. I'm just keeping an eye on the time but I think I should be fine. And I'm coming to the end so here we could say that yes we had the similar feeding patterns have similar impacts. And now I will come to my conclusion. So here I just wanted to, or like if I have some sort of messages for you to keep after this talk is that in our case, these parties had a relatively broad dietary niche, dietary niche breath, but that didn't didn't mean that they didn't have very distinct feeding patterns and predatory impact so you can feed on different on quite a broad community but still have a distinct impact in your system. And I do think so if you, especially, we also just heard it from the from previous talk that soil is very heterogeneous, and you actually better do not specialize in the system otherwise you will probably starve quite quickly. So that does completely make sense that sort of predatory parties would not specialize too much in the system. And another point that I found quite important and interesting is I think that we can really use artificial system like our feeding pattern essay to study microbial life strategies of diverse parties so we can use artificial system and then relate them to more complex systems. And this can really help us to go further and understand better all these different systems. So, with that I think I will come to the acknowledgement otherwise I'm running out of time. Just want to thanks to people of the group I was during my PhD, the ecology and biodiversity group at the university. And all this work was supported by a envio grant from the Netherlands. Special thanks of course to all the co author of this paper, and to the organizer of this webinar. And so to my current organization who is allowing me to give this nice presentation, and to all of you for your attention. Thanks. Thank you very much, Natalie. Very interesting. I also again have a bunch of questions but I'm going to save these for later and of course everybody else please type in your question to the q amp a and we'll get back to them so no, really interesting in terms of how bacteria are our food. So, our next speaker is Eric right from the University of Pittsburgh and looking at model communities and multi stable bacterial communities and how they are infected by the conditions of the system so. Eric, okay, I see you got your camera back on and hopefully you can get your presentation up and we'll get you started. Okay, great. Thank you. Let's see here. This is real quick. See if that works. Okay, okay looks good. Thank you. So, hello everyone, my name is Eric right I'm from the Department of biomedical informatics University of Pittsburgh. I'll be presenting to you today some work I did actually in the laboratory of Colleen Betsy again at the University of Wisconsin with postdoc name or sorry undergraduate named ravina group that who's now a PhD student at Northwestern University in Chicago, but the main stars of our show are actually these guys the streptomyces and if you have never heard of the streptomyces they're very cool genius of bacteria as you can see incredibly morphologically diverse. All sorts of different colors, because they are well known for producing antibiotics, and they are ubiquitous in soil. And if you take a grain of soil that is microscopic, you can get hundreds of streptomyces colonies on selective medium as you see here. And this is basically just a starch and agar medium they're happy to grow on that and even just on starch, and that's because they're sacrifice, and they break down decaying organic matter and nature that's their sort of role and they so they can grow on very very nutrient. They're well known for making antibiotics each strain produces on the order of 50 different national products and we can see that from their genome but many of these are actually silent in the laboratory so we're unable to express them or haven't found the conditions to do so. What's interesting about them is that there's very little overlap between strains so if you took two different streptomyces from what I just showed you from the same soil sample you'd see from their boss the gene clusters they encode there's there's very few antibiotics that they actually would share. And because of that they have this enormous diversity of antibiotics they produce about a little over half of the antibiotics that we use in the clinic right now so if you take an antibiotic for it there's a good chance that came from a streptomyces or a fun guy originally. Not only that but because they are also the major antibiotic producers they're also the major source of resistance right because they can't kill themselves with their own antibiotics so they develop these resistance mechanisms that happen to be sort of the lowest cost resistance mechanisms that are possible. And because of that they are both the producers of antibiotics that we then use in the clinic but they are also the source of many of the resistance mechanisms that you see in the clinic. But the reason I'm talking to you about them here today is because they have a very cool life cycle and we've been able to link some of the results that we've seen to to their life cycle. You can think of that life cycle as starting as a free sport, and this sport if it lands in the right condition will germinate. And one of the things that Cleans lab did a lot of including myself when I was there is we studied this germination process because it's heavily regulated. Once they germinate they grow down into the substrate and actually if you if they germinate an auger you can actually cut the auger with a razor blade and you can see that they grow down into it quite deeply sometimes a centimeter. And after some period of time, they make some developmental switch where they decide to grow back out of the auger into what's called aeromycelium. So the switch from a filamentous growth strategy to developing spores so the filaments that contain about 40 different genomes that are only loosely compartmentalized and there's not a ton known about how they're compartmentalized but there's newer work showing that they are. They they septate, and once they do they become individual spores, and they can disperse in the environment okay so this this step of germination and the step of speculation are both heavily regulated and very important to their life cycle. What we were interested in doing is understanding how streptomyces would assemble any natural habitat, and we imagine them as first starting off with these free spores that were dispersed to some new food source and because they're there might be fights that might be a leaf or it might be, you know, a piece of wood that fell on the forest floor, and now it's ready to be decayed. And so they arrive there presumably quasi randomly, and then over time, they grow, and the environment will select for which ones are capable of growing in that particular niche and also they can inhibit each other, because of all these different new encode, and that leaves behind some final community that then presumably goes on to inhabit new fresh habitats when another leaf falls on the forest floor. So, you can imagine this process is happening all the time and we want to understand it and understand how communities assemble in nature. And really there's two ways that people have studied microbes, most of the time and that is through natural systems, and which of course are extremely complex, difficult to manipulate their couple to other complex systems, and it's very difficult to measure interactions in nature, although not impossible it's it's not as easy to manipulate them. Or pure cultures. And most microbiologists you know study individual organisms E. Coli K 12 or something like this and a petri dish and so this doesn't tell you that much about their community context. And so in order to sort of bridge this divide between natural systems which are too complicated and couple the other complex systems. And pure cultures which are too simplistic to study communities, we wanted to develop synthetic communities and there's basically two ways that people go about doing this. One is through a top down approach where you take a natural system and you constrain it by allowing it to grow on some resource in the laboratory typically or sometimes in people's bathtubs there's all sorts of different things that have been done, and that reduces the community to some alternative community that is a subset of the original membership and creates a synthetic community some of which are sustainable, or pure cultures in which you start with a few organisms and you would try to assemble them bottom up and the challenge of this is that lots of times you run into competitive exclusions so that you only end up with one organism in the end. So this is a pure culture approach and we wanted to understand basically everything we could about a small subset of organisms. So we started off with a panel of strains which I'll be representing as these sort of circles that are colored to represent different colors of strains. We measured a bunch of different attributes about them their growth rate their yield in terms of number of spores they could produce or biomass they could produce their lag time things like this. So we collected their competitive ability, whether they're able to invade each other which I'll spend most of my time talking about today, and also where they can inhibit each other and we want to link their inhibition to their ability to invade if if there was any sort of there we want to understand that. And then we sought to capture multi species abundance is over time so if you put a bunch of different assemblage of community members together what happens if you just let them sort of grow together. What do you observe is their dynamics and then we wanted to be able to model this so ideally if we could understand the system then we'd be able to construct models that could sort of recapitulate those dynamics. To do that we ran this experiment that was rather simple and when I first started this work I was somewhat amazed that this isn't something that had been done I spent a lot of time trying to figure out if people have done it in like the 60s or something and I just wasn't able to to find such a thing So, so we did a relatively simple experiment where we took a small fraction of one strain about one 1000th of the community, and we put it with another, we call them the invader in the resident so the resident is the one of this the majority of the of the community. And we embed them into auger within a test tube and that test tube is basically lined with auger I'll show you how that process works in a second. They grow inside the auger and they go to the top and they form aerial mycelium and then they sporty latest part of their life cycle, then we would harvest and freeze these spores and repeat this process. And the neat thing about this is we didn't need to do this too many times, actually the dynamics played out very quickly in almost all cases and within about three propagation cycles you would observe a final outcome that we were able to confirm in many cases that it would last for basically forever because you ended up with just one strain or coexistence or something of this sort. So, in our first experiment here which was published in nature communication a few years ago. We use DNA sequencing to tell the population abundances so we got relative abundance out of that we sequence the RPO be the RNA polymerase subunit beta gene, and we observe the relative abundance over time starting from some strain, which was shown here in green and then we looked at whether minority strain could invade it so we call this the invader. And you can imagine that there's three possible outcomes you could have survival of the fittest so that is if the purple strain can dominate the green strain then if you put the purple at low abundance it can invade if you put the purple high abundance it can be found. And so we would say there's a hierarchy between these two strains. And this is sort of what we went into this, expecting to see the majority of this. And then the next we outcome you might imagine the survival of everybody which is, if you have two strains that can coexist together then either one as a minority would eventually reach some equilibrium. And I represent this with a double headed arrow to say they can both invade each other. And a third possibility is that you can imagine, we could observe survival of the common, which is that the more common strain is the one to take over the community. And this results in by stability, in that if you start with the green strain as then Vader, it is able to or it loses ground and if you start with the purple strain as the invader then it loses ground and so whoever is the most common ends up winning and we represent this with a dotted inward double headed arrow. So we expected to see mostly survival of the fittest but these are the three possible outcomes you can imagine. And then, once you have that for a panel of strains you can construct a pair wise invasion network and that pair wise invasion network might be hierarchical, where you have one strain that sort of rules them all. And one strand at the bottom that's beaten by all of them at least in your laboratory community. And you would imagine that if you put all these strains together all four of them from knowing the pair wise data you'd be able to predict that the purple strain will win. And you could also imagine a network where you had some arrows that went against the hierarchy. And so you might have loops in your network. And these loops are thought to maintain diversity and natural ecosystems. Or you could imagine that the notion of hierarchy entirely does not apply the system because you have arrows pointed all different directions, and of all different types. And, and so you can't arrange these particular strains at least in this setting into a hierarchy where one is better than the other. And so to do this we started with a panel of 18 different strutomyces here's how they all different all the ones different look. All of them are strutomyces except one strain which is strain number one which was an amycolatopsis which is a related genus, and we threw that in there because it was, we're interested in what happens if we put in a non strutomyces. And we had two replicates of the strain and all the others. We had one starting from each initial condition. So, we also included most of these I should mention are from the same soil sample, except strain 13 which is strutomyces celicolor, and this strain is well known to strutomyces people, because it's sort of the canonical strutomyces. So, this is how we ran the experiment, we injected them in molten auger and we built this contraption that basically blows air over the molten auger while it solidifies and it rotates at about 1800 rpm and actually the, the motor here is driving this is our lab technicians that had broken down. So we made a motor out of it we had these o-rings and this was quite a system trying to do this in high throughput and we had hundreds and hundreds and hundreds of these tubes every few days so so it was a lot of work to get this this working correctly and kind of anything. Once the tubes solidify this is what a negative control looks like so there's nothing in here except auger and so it looks more or less clear. And this is what it looks like once the bacteria in it so you see a lot of bacteria growing inside and this point they've correlated and then we would inject in a glycerol stock for freezing and harvest the strains. So, this is the outcome of this experiment. We measured all possible pairwise invasions between every combination of two strains, and you can see the invader here on the x-axis the resonator here on the y-axis, and if there's an invasion is shown in red. And what you immediately notice is we had a lot of invasions. So we categorized those into our three different categories you can have by stability so here Strain 4 was the invader was not able to invade Strain 9 and when Strain 9 was the invader was not able to invade Strain 4. So you can see that by looking at cells that are symmetrical across the diagonal here. We also can see that cases of coexistence so for example Strain 1 is able to invade Strain 17 Strain 17 is able to invade Strain 1. And we can see cases of hierarchy so Strain 12 is able to invade Strain 14 and Strain 14 is not able to invade Strain 12 so we have hierarchy between those strains where 12 is presumably more competitive or better in this particular system. When we broke those down into our different possible types, our three possibilities, we saw that the by stable pairs were actually much much much more common than we originally expected. And this sort of blew our mind and was sort of our first major result that wow we were onto something here this is neat we expected a lot of hierarchy and said we saw a lot of by stability. And actually I'm not going to have time to talk about today but we link this by stability to their ability to produce antibiotics so their natural products that their ability to inhibit each other results in this by stability and that's that's one thing if you're interested you can look at the paper and learn more about that. But if we take all these and then we try to construct a hierarchy of strains and say who's higher in this hierarchy than than the other strains, we see that we end up with a hierarchy like this and this was based off of an algorithm that was done to form hierarchy and social networks. And that we sort of copied to to figure out whether our data shows a hierarchy, and the surprising thing to us was that a lot of strains were at the same level of the hierarchy so so somewhat flat. There was no winner there were actually six different winners. And if we look at the if we overlay the invasion results on top of that here's the arrows of who can invade who we see that these these strains at the top of the hierarchy can't invade each other. And mostly, as you would expect for a hierarchy. Most of the strains can inhibit the, sorry, most of the strains at the top inhibit the ones at home. So we actually look at which arrows go against the direction of hierarchy where we see invasions that we don't expect that are not conforming with the sort of competitive hierarchy that we develop there's a few of them and almost all of those are with strain one which was able to coexist with most other strains. And this is the strain if you recall that was an amiculotopsis so we saw a lot of coexistence with the amiculotopsis and almost all the strepomyces were not able to coexist with each other. So we overlay by stability on the network you can see there's a lot of it. And actually these by stable pairs are enriched for inhibitions. So this was a really interesting thing and ended up being a more surprising result than what we expected when we set up. But there were some limitations to this work, which is why we did another study. So the first thing is that we started this one to 1000 initial ratio which is somewhat extreme. We don't we're not able from that to tell exactly where the boundary of by stability is this is also known as the super ticks. Secondly, it's difficult to derive dynamics from the sequencing data these their relative abundances and we would ideally like absolute abundances for modeling. And then thirdly, we never tried more than two strain communities so we couldn't answer the question of whether the by stability that we observe propagated multi stability. What we did is we for the fins microbiology ecology paper. We reduced the number of strains, and we looked at more starting conditions. But we did it in a special way and that because we wanted absolute abundances, and we couldn't figure out a way to get that from sequencing without all the inherent sequencing bias and PCR bias and whatnot. We had to count the strains. So this was quite a laborious experiment I wish I had taken pictures of myself with all the petri dishes because we had piles and piles and piles and piles of petri dishes so many boxes of petri dishes trying to count these strains at different concentrations to get absolute abundances. And we chose these particular five strains, because the local autopsies showed some interesting behavior strain two was pink, which I thought was neat. So strain three turns blue this deep blue strain for sort of a salmon color and strain five is great. And so we kept their naming the same as our original paper so we have one two 131517. And we chose them because they're the exhibited interesting behaviors. In our first study these three were at the top of the hierarchy and the top level there with the six winners. They're totally different colors from each other and so we wanted to know. We want we were able to look at them on petri dishes and pull them apart from each other. So we did all possible pairs, all possible triples, and a subset of quadruples. And then we constructed, once again the invasion network. And what you can see is starting from different initial conditions those are the points here. So the pairs of strains the strain one insane strength 17. If we look at those starting from different initial conditions more of one less of 17 more of 17 more of one. And here in the middle, an equal amount of 17 one, our final outcome after three propagation cycles as everybody moves towards this sort of central attractor and that is between strain one 17 so there's there's both strain still present and strain two are able to co-exist. We saw that if you mix together one in 15, you almost always get 15 in the end after three propagation cycles. And so these two strange show a competitive hierarchy where strain 15 is able to dominate strain one. So we see the point of by stability between 15 and 17 you can see is right here. So this separates the region where if you have slightly more of strain 17 you fall towards 17 is your final condition if you have slightly more of 15 you fall towards 15 is your final condition. So we did this for both the pairs and which we had already done previously but now we have more more starting conditions. We also did this for points inside the triangle, which are showing you if we put together three strains, and you can see that this region of by stability propagates throughout the entire triangle so so here in the center. You have a point where everybody is at equal abundance. And actually at that point the two replicates diverge, which was really quite amazing. We actually sampled this densely enough to find that our replicates would diverge to different final solutions. That was pretty remarkable to us and you see that in a couple cases here where they started the same initial condition we have two replicates and they diverge to different final conditions but this by stability propagated to multi stability. But you can still see the basins of attraction here they're, they're quite clear. And it's really neat to see these, you know, in in high definition, given that this is a well studied thing in ecology these basins of attraction a lot of it's done through And if we do this for all possible triangles between our five different strains here's the outcome. And you can see that we have a bunch of different attractors I think here there's let's see 12345 different attractors, and there's a large region of by stability between several the strings. This was really cool and this showed that the by stability that we observed and in pair wise communities propagated to multi species communities. And so we want to know does this keep going to four species, we did that and these are points now inside the triangles and so here in the in the middle of the pyramid we're starting from four different strains. And you can see that once again we have the different attractors, and in the final solution is quite different depending on just slight changes in the initial condition. And actually at this one point here, you can see they diverge to different final solutions. And in case you're a bit skeptical of this and thinking that this must be crazy or artifactual, you can actually see it in the tubes is one of the amazing things so we have these test tubes they're embedded with our communities and this is after three cycles and this is my nomenclature so we have strength and we have community to a six a into a six b these are two different replicates starting from strains 131517 and they diverge after three cycles they remember they start from the exact same volume it's literally mixed together. Take a pipette take 100 microliters of it out and inject it in two different tubes and that's how they begin and they end up with dramatically different communities here you can see strain 13 strain one is still present and low abundance here. And here you can see strain 17 it's sort of salmon color there, and you can see strain one and slightly higher abundance there so so this is a remarkable. confirmation of what we observe from that central point starting with four for communities and and we were able to see this in the tube so it was really neat to be able to watch this experiment not just through sequencing the deal see it happen. And then finally, we did quite a large modeling study where we, we, which only some of which we publish which is we publish the lock of Voltaire part of it but we also tried to model it in other ways and we were able to model it better than but we never ended up publishing that because we had constructed all sorts of crazy models and we weren't sure anyone else would would be interested in them but starting from each of the different initial conditions, you can see lock of Voltaire is able to qualitatively capture most of the outcomes and so the points here are the real data the lines here lock of Voltaire predictions, and it's colored based off of how bad the prediction is. And notice that the the axis is in log space so this is the finals or abundance after some number of propagation cycles. And in some cases it did very poorly right so here it thinks blue wins actually green wins and green, it thinks green goes away but green one so and you end up finding that these are near the boundaries between the basins of attraction. And so it's an optimized lock of Voltaire system which contains many parameters you have a parameter a rate parameter for each strain for its individual growth rate and then you have interaction parameters for every pair strain so it's a highly parameterized model is still not able to fully capitulate the dynamics, and not only that but I guess another point to emphasize is that the quantitative dynamics are way off so so take a look here at this strain this is strain one so our emicolatopsis grown by itself after one propagation cycle shoots to tend to the eighth spores per mil and instead we have our sorry CFU per tube, and instead here you see that lock of Voltaire is expecting to slow rise and eventual plateauing. And so it's just orders of magnitude off especially in the first propagation cycles and we were able to construct models that fit this data better but we ultimately didn't publish them because they were a bit, a bit custom and but they sometimes have fewer parameters. And so Voltaire, which is well known to ecologists did okay, but not superbly on this particular data. With that I'd like to acknowledge our different funding sources and that funded this particular project, and also if anyone's interested I have my own group now and you can look us up we study antibiotics so an easier problem in some kind of biotic resistance then modeling microbial communities which is, as you can see here quite, quite complicated and I think that's really the takeaway from all this is that streptomyces, even in small systems with just four or five strains are incredibly complex so you can imagine the real dynamics when you have thousands of strains and a true microbial community in nature with a lot more coupled systems like the weather and other things that are interacting with these strains. So you have incredibly complicated dynamics, and that should be no surprise when you can get such complex dynamics, even from smaller systems. Thank you. Thank you Eric. This is great and thank you also Natalie and Damian, overall, really interesting and I really started to first think of the quip from Jim Proster from a few years back because why do we have this ridiculous complexity of microorganisms and soils and other systems and it's interesting to look at yeah what are some of the drivers, I mean from competition between strains, close related to each other, the predation, as well as then of course, metabolic impacts of really, really interesting and very different approaches so yeah we'll, we have time some discussion, there are a few questions here already in the Q&A and so we have about 15 minutes to go through some of them, probably not all. So yeah Natalie and Damian if you'll also join we can then start to start the discussion. And what I was wondering, yeah this is one from Maria Martino, this is actually going back to Damian, this question of oxotrophs, are they cheaters? What's going on here if you think of this aspect of feeding, are they giving back to the organism that is providing the amino acid or something else? I was looking at your question in the Q&A Maria and trying to think how I would best respond to it actually. Unfortunately I didn't come to a good conclusion there. I think often it can go several different ways either you can have a microbe that over time has like streamlined its genome and it's lost so much of this extra stuff that it needs that yes it probably does become a cheater and then it doesn't provide anything useful to the community and then that sort of one outcome. But I think another outcome is possible, I like this hypothesis it's called the division of labour hypothesis whereby two different organisms they've lost complementary growth factors and then they have to sort of like live together in a more mutualistic sort of scenario. I don't know in the real world how frequent one of those other sort of possibilities might be. If I can trick people into giving me money I would like to answer that though. Oh and actually and there's a follow up this is actually from yeah Petra Stefan on the deep biosphere and this is again going into severe nutrient limitations. I mean you're looking at systems and soils where there's often a plentiful amount of nutrients but as you go to more limiting systems you see this mutual cross feeding becoming really more and more important and what's the evidence of that? I'm not entirely certain about the very deep biosphere. I have seen some papers where there are these this phylum Patissi bacteria that seem to make up quite a large proportion of communities from some of these sort of ground water or deep biosphere environments and they always seem to have suspiciously tiny genomes so very much these might be organisms that live that sort of cheetah lifestyle and then they're forced to either find a good partner that they have to live with in the deep biosphere or maybe they just get a lot of sort of like ground water or mineral flow from more rich active sort of environments and if those substrates can make it down to the deep biosphere maybe they're just relying on that but I probably don't think that's a very safe way to live it would be very risky indeed. So, let's see Satya Singh has a question regarding predators. I mean, what's the basis for predators eating specific bacteria I mean I guess surface recognition or so on. And I wanted to add on to this question, thinking of it from the perspective of the bacteria. Are there some that just taste bad and therefore evade predation versus in terms of what is the feeding preference of the of the protest. Thanks for the question that's a nice one. Yeah, so for there are like different ways of why produce would feed on one or the other bacteria so indeed there is some surface surface recognition for some of them. So some parties are able to recognize some bacteria and they then decide they want to feed on them or not. We expect that there is a system stoichiometry importance there but we don't know exactly which bacteria are preferred or not. And so, I'm coming then next to the other question I'm saying back to you all the time but I just want to make clear that predatory parties are not only feeding on bacteria. This is just the most common, the mostly studied organism but parties can also feed on other organism including other parties, just that not people think now that products are only feeding on bacteria that would be wrong. And if they taste bad, I guess so because some of them do produce some toxins that can not only taste bad but be deadly so some parties are so able to some to yeah to sense that and then they would avoid this toxic back to I guess that answer the questions hopefully. So wonder that whether that also links with antibiotic production or some of these, I mean, for example, streptomyces and so on and there. At least we had some we have another paper on that actually and we could find some good correlation with this that's a, if a bacteria was able to it was with Pseudomonas but if the bacteria is able to produce antibiotics, it has bigger chances to not be eaten by the Okay, yeah, I'm just noticing another question from Lily young, I think it's mainly to Eric, and I think she's asking to complicate your experimental setup even more in terms of okay you have multiple strains competing but this is one of showing Yeah, even two strains depending on the concentration of the substrate, the competition outcome will be different. So again you have scavengers and so on so how do you see this, adding to your multi strain competition if you now have nutrient limitations. Yeah, so we didn't test that. It's definitely a major factor. The medium that we used was starch media it contains a few salts and primarily I think it's one gram starch per liter. So it's highly reduced very few things will grow on it. And we never tested other types of media. We're, I mean we did, but we didn't publish any of it, and we know that it has a big impact. So you can change the, you can change the outcomes of competitions by by changing the media for sure. So, yes, more more dimension since your model. This is going to be be interesting and how to keep track of all of this. Let's see there was also regarding the predation. So Natalie you looked at protest and they're feeding of bacteria what about the other way around. I mean I know they're nematotrapping fungi. But what about bacterial predation of the protests. That's potentially a good one. I don't know any research about that I wouldn't. Well, I mean we know that some bacteria can kill the parties via toxin production, but that they would then feed on the protest. I wouldn't know that no. We see some order interaction with other system that there are some parties they can feed on nematodes for instance and nematodes get fit on parties but some parties can also feed on nematodes. But otherwise, yeah, I don't have something to research I guess to see. Yeah, I mean definitely under the microscope I've seen is have done once a protest dies, you have a zoo of bacteria accumulating around there and probably feeding on on the substrates but already specifically killing it they're just waiting for for that big protest to to collapse. That's a good question because you also have endophyte so it's some bacteria are so leaving inside the parties so maybe they could so try to kill the parties from inside to feed on it. I don't know. I really I don't know. Many research on that so but it's a good one. Okay, I'm starting to are there others here that you've noticed that you want to jump in that. I would just want to say one more thing because I saw somewhere a question about sort of sort of extremo feel protest. And there are some parties that can only there are some silly it's something that can only feed on spores of fungi and they really have only the specific enzyme for that so this is so does exist so we do have some very highly specialized predatory it goes really I think in all the direction so for the parties in this diverse group that just wanted to add on that. Okay yeah and then one is of course thinking about the interactions even beyond our systems. This is from Connie smaller on fungal high feed it of course often gets connected with specific bacteria that are growing on that what about strip to my sees and so on are there other bacteria that are typically associated with them I mean how would that fit. Yeah, so I don't know of a nutrient exchange that's been well studied between strip to my sees and other bacteria and I may be remiss and not remembering. So, I'm not 100% sure I don't know of any correlations between other organisms and strip to my sees. I can tell you that one rather interesting thing that is it's still unknown, why it is the case is that both fungi and strip to my sees are our main sources of natural products right. And, and so why would that be they both inhabit a similar niche is sacrifice they both compete with each other. They both grow filamentously, and they both have long linear genomes. And so, so it's quite a lot of, you know, associations, but, but to my knowledge no one, no one has a certain answer as to why that is the case but it's thought that the strip to my sees and fun guy because they are the sort of keystone species in degrading organic matter, they digested extra cellularly. And because of that need to protect their food because they're basically breaking down these large polymers and the things that get inside cells and once they've done that work. They now can have competitors come in and take the outcome of that without having to do any of the work to do that biology or the chemistry sorry. And so they, it's thought that that's the reason the main reason why they're antibiotic producers and they have so many different natural products is that they're protecting their food source basically because they have to digested extra cellularly. The main thing is you do all the work and then somebody else eats the sugars that you've just released. So, actually, I thinking of that this is for for Damian on on also in terms of the interactions between the provider and the oxytrope and the heterogeneity that we then have in soil systems. And to see in terms of associations again if we can really analyze a soil particle to look at, do we have the provider and the oxytrope actually growing as a micro colony in a soil particle. Or how would that I mean again you, if they're distances too far for diffusion, you're not going to get that benefit. Yeah, absolutely. There's a fantastic paper from Renault and London. I think that came out in a few years ago 2018 perhaps. Look specifically at modeling the spatial separation and what's realistic in like a soil environment, and they were some of their models were predicting that within a sort of 1020 micron sort of radius. You really only have like maybe 100 cells, maybe 1000 cells but then only probably like, maybe like, I think it was something like maximum of 10 different sort of like you know taxa at the ASV level that could actually sort of like you know be present. So, yes, I think it is possible that you could have actual sort of like you know cross feeding of organisms that are growing like you know together probably in a buyer form of course because of the environment that's how these dudes sort of like you know coexist. But yes the spatial separation is so large that our idea of looking at sort of like these communities as 6000 different sort of ASVs and 250 milligrams. That's just like not realistic, I don't think. Yeah, and really that the interactions are far fewer than what we think based upon our integrated analysis of a whole soil sample. Okay, no. Thank you, I think we are out of time and have to wrap up so Eric Natalie Damian I want to thank you so much for joining us really interesting to hear you talk about the work. And again, there were a few questions, some of them technical details that we didn't get to. So for those please read the papers that you have the links to them and some of, I think the answers will be found also there and Again, thank you all who joined us from around the world again it's a wonderful audience I've several good colleagues in the audience so and for others who I don't yet know so again thank you for joining us in this webinar and again thank you Eric Natalie and Damian and see you next time.