 and and we can we can read them. So Chiara are you there? Hello, good afternoon. Hi Chiara, nice to see you, can can you share the screen? Sure, so do you see my presentation? Yes, it works, very good. Okay, perfect. So let's move to the next presentation by Chiara Poletto and understanding the impact of cross-front dynamics on pathogen diversity. Thank you very much. Thank you. So thank you so much for the invitation. Yes, my name is Chiara Poletto and I'm a researcher at Team Ceramic Paris. And okay, I think I will just introduce myself at work on mathematical modeling on infection transmission from human to human. So it seems to me that it is a very different scale with respect to what I saw so far during these talks. But I'm not an expert of molecular epidemiology or classmates, but I hope that in terms of let's say, approaches or perspective, you will find the method interesting. In particular, I would much appreciate your feedback in such that what I'm going to present you is some theoretical works, let's say on on some interaction, ecological pathogen, ecological aspect of pathogen spreading. And now we are continuing trying to figure out the implication of our results in practices. Okay, so let me start. I'm interested in the structure of human to human contacts, contacts that may get a transmission of pathogens, and in particular this contact network. We know that as a critical role on the risk of an epidemic at the population level and on the epidemic trajectory. And it's also important for the design intervention. However, the majority of studies. There is a strong attention on like epidemic of a single pathogen spreading on the top of the nest, this network. But actually it's not true and clearly there is, there is, there are many pathogens circulating at one time or also we may be interested in considering the structuring of a pathogen. In in in its multiple strains, because this may have an important role concerning emergence dominance coexistence or ecological patterns patterns in pathogens and we know what clearly you know more than me how this is important in terms of antimicrobial resistance but also I'm interested in other infection like influenza infection like covid and this this is important for a vaccine design or for anticipating the trajectory of epidemics. So, if we are interested in in the drivers of emergence of new strain for instance or dominance coexistence patterns among different strains we know that there are multiple factors we need to pay attention on, we need to pay attention of course on treatment or interaction mechanism that may act, act among strains so we know that strains can compete or cooperate or they, maybe they don't interact too much it's depend on the pathogen. Actually, I'm interested on on a higher level and in particular, actually, how the population behave the human population behave and interact may affect like the ecological dynamics between strain given that strain spread on the top of this population. And in particular, my specific focus would be on contacts so the structure of the population network and this specific subfield let's say it's a field in which there is some interest that I'm reporting here some, some publications. So the point is that if they consider to, let's say let's imagine two pathogens or two strains to spread the agents and let's say with different rates like different transmissibility or different, let's say duration and infection. And the typical question is which one is favorite so which one will dominate. Well, well, actually, there are several studies, each of these focusing on different aspects, but actually in common the point is that the answer to this question so which will dominate while the answer is depends on the structure of the population. So one pathogen may nominate in a certain population while another may be favorite if the population has other characteristics in terms of contacts in terms of behavior. And okay, we did some work on this direction and I want to let's say we'll discuss in particular for sure one publication maybe if I have time I will also discuss a second one. Okay, so let me start with the motivation of our work. I stress that it was a theoretical work, but it was motivated by, by the spread of Saferococcus arrows in hospital settings. So hospital settings have peculiar characteristics in terms of population belay behavior. There are a small number of individuals, relatively small so for sure less than a city or a country we have 10s under the units of individual confined space when we have three individuals a stochastic effect became become very important. And then we have a high turnover. I mean patient and turn the hospital spend a certain number of days and then get out and this turnover of course affected the interaction between people in the hospital but also is responsible for exchanging bacteria or pathogens in general with the outside of the hospital. And then we have a structural feature of the network like you have vards, and so people tend to interact more within the same world, but still we have exchange of people and bacteria across vards. And we know that there are contacted origin it is what does it mean that some people are more active so make more contacts than other people. If we think about the hospital of course we know that alcohol workers make a lot of contacts with all patients and among among each other, why patient is that move less in general make less contacts so alcohol workers cover an important role on the propagation of bacteria in hospitals. And these all these elements are recognized to be central in the circulation of bacteria infection in hospitals. And, and, but still, I mean, the study of this element was quite boosted recently, at least an important contribution to these was the use of network data to to characterize the pattern this pattern of interaction and in particular we have data on the exchange of patients across, across vards or across hospital, all the data of face to face interaction. So, while I would go directly to this slide so we present this study that is the individual based investigation of resistance dissemination this is the hybrid study. The pleasure to to to work with this data even if I was not involved in the project this project is a collaboration between different institutions in France like the pastoral in sermon. And what we did was a simultaneous so it was a long term, a facility, structured in five words, and for a duration of four months they monitored, monitored the face to face contacts, this can be done with the RFID technology means that it's a sort of tag that you wear. So there's people to wear this tag, and it's a highly sensitive in the sense that you can detect only face to face interaction because it's sort of signal that you send from attack to another. And the signal is weak so if people are far from the other, you don't see any any signal. And so you can detect really the face to face or also physical contacts, if you make the signal very, very, very weak. So these are the kind of interaction that may be considered a proxy of the interaction relevant for the transmission of of of nosocomial infection. But what he was interested is that in parallel with the collection of this contact data, they also collected the colonization data and colonization status. And so we, we clean as a swaps. And, and these, these, these swaps were subtype. So we could, we can observe a plot like this one in which we can see, we can see the different strain spreading week by week. And we can see that there is a dominant strain, other strain maybe are less, less prevalent. Okay, so that was, I mean, all of these motivated a theoretical question. The question is how the structure of the network contact network determine the plot that you see on the right so the ecology is seen at this level between the different bacteria strain. So the ecology here can be can be characterized in terms of overall prevalence, but also in terms of richness so that means how many strain and see, in terms of a Buddha so which strain in particular we can define also we can also analyze the indicator of evenness or dominance. So do we have that the strain dominating dominate a lot or do we have a steady situation which all strains are more or less equally distributed among infected cases colonized case in this case. So we can use an index using ecology that is called the Parker that is simply the portion of the dominant strain over over all strains. So if this is a high if this goes close to one clearly have a situation in which we have dominance so this strain is dominating. But if this value is low means that we have a equal partition amongst trains. Okay, so we wanted to understand how contact heterogeneities affect these ecological indicators and we did these true numerical simulations meaning that we can simulate a situation in which I mean we can simulate a network of of contacts that is say they produce the same property of contact networks in in the hospital setting and particular what what we design is a dynamical network in which every day not here I'm talking about notes about notes represent individuals. So we know that individuals every day make different connection or every hour every second so in this case connection changing time and we have also individuals entering the system so entering in the hospital and carrying with new bacteria or new pathogen in general. So, every time an individual entering the hospital many parties is that it brings a new a different strain. So in principle we have an infinite population of strains and we consider an ultra hypothesis in the sense that we made the hypothesis that strains are identical in the way, let's say in the epidemiological trade so transmissibility and duration of time so we speak about given that I'm a theoretical model that sometimes I speak about the infection or colonization I mean, in reality this represents more the colonization of of the field coco hours. And, okay, so if we compare it's an important point and it's a genius network with an homogeneous network what doesn't mean the genius network some notes are more acting the others. So they make more connection and these are present really situation like medical doctors and and like patient, why we compare this with the situation which we have an homogeneous network, which all individuals are more or less equal in the way they behave. Okay, doing our numerical simulation find something that more or less seems the picture of the data in the hospital so the stochastic dynamic of multiple security strain, but actually if we look at the indicators, we can see that according to the network, we have different outcomes to decide ecological dynamics is is different. So we can see on the left that the prevalence doesn't change so much. So we have that the homogeneous network. Here on the x axis I'm firing the trust me stability so I'm sprawling the political transmissibility of this pathogen. So increasing the transmissibility of course the prevalence increases and the homogeneous and the genius network prevalence values levels are more or less similar. So if I have a highly genius network so gamma represented the energy lower is gamma higher is it. So if I have very serious network I can see that the prevalence is like lower but the difference is not so big. Actually, I can see that the difference is bigger in terms of richness. So if I have an indigenous network and I have lower levels of richness, so I am more to accommodate a smaller, a smaller number of pathogens and okay, it's interesting if we still on on the right. The the burger Parker so the burger Parker quantify the the dominance that we can see that in the indigenous network we have higher level of dominance. So. Okay, so this is this which are to understand why the network may change in this way the ecological dynamics between strains. And what we find is that we find an explanation this maybe understood in terms of the property of the network itself. So there is a nice study by 11 telecollaborators that was published in 2015 that explained this mechanism, why in the genius network prevents. Let's say, in the emergence of new pathogens new strain. That's because highly active nodes tend to be in general always colonized because they are more easy to reach by infection. To colonize making the hypothesis of mutual exclusion. It's harder. I mean, for a for a new strain to to to to percolate in the network. And, but at the same time, and here is the plot that I'm showing the bottom, if a strain manages somehow to colonize a highly active individual it was spread through the network. And this is the so called super spreading event. So we will create situation of dominance. So, we try to analyze this with respect to the hospital network. So, as I told you at the beginning I was showing these these this plot in which we see the different strain circulating the data. So we have to check the burger park index, we can see, I mean, a certain level of the burger park index. And so we try basically to make simulation actual network data that was measured and compare the simulation with a randomization of the network. So we through randomized the network with destroy basically the properties of the network in particular the technologies. And what what what we can see is that the simulation in the actual network. This is the blue band that I'm showing the confidence with interval show a level of burger Parker that is more compact compatible with the data. And the simulation in the randomized network show a burger Parker that is lower. And it's less compatible with the data so this means that the heterogeneities in the network may, let's say, somehow explain part of the dominance of serving the data. And at least they contribute in explaining the dominance of serving the data. So I concluded this this part of my talk. And so we tried to do to use to adapt an ecological perspective to characterize chain population. And what we found is that contact heterogeneities in the introduction of strain from the from the outside and so these reviews the richness of an ecological system. But at the same time, the few as a successful strain are amplified and create super spending events so we have a situation of more dominance. So the model so far is very simple. In the future we want to account for others as moralistic aspect like, for instance, maybe not a complete mutual exclusion, or also differences in epidemiological trades. Okay, I think that I finished my time I was having another project my baby. I don't know if I would have the time to, to explain it so I just go directly to a knowledge mentor so I wish to thank my micro actor. And the worker presented was published before COVID pandemic. And here is the reference. Thank you so much for your attention. Thank you. Yes. So, any comment or any question. So, otherwise, we can move questions and answers to. I have a question myself so how important is this assumption of exclusion between pathogens. I mean, you say that when one individual is affected by a strain it cannot be affected by another strain. So how is this. So are you looking at a single say infection or a single pathogen and multiple strains and and you say that in one pathogen in one patient that is only one dominant strain. Yeah, so in my case I make the potages of motorist cruise if I could organize by one strain I cannot take the other. And this is, I think reasonable and now that people may be in the audience know that more than it depends a lot on the bacteria we are considering so first of all, I think this mechanism is really the key. Well, I don't think it's necessary that is a total mutual exclusion to obtain these results. But this kind of at least a partial mutual exclusion I think it's the key for for creating these these ecological mechanisms I was showing today. So I think this is in which instead there is sort of a you relax a bit this assumption and the and I know that the mechanism you may find that quite different. I think this is something we will be. I mean, we should explore more in the future. Thank you. Thank you. Thank you very much.