 So, good morning, good afternoon or good evening, depending of where you are joining us today. We have the pleasure of having Professor Lance Price from George Washington University today as a speaker in the Knowledge Dissemination Dialogue of this month at FAO. This is a webinar organized by the FAO AMR Working Group and the FAO Sustainable Livestock Technical Network, always happening on the third Thursday of the month, around lunchtime here in Central Europe. A topic that Professor Lance will speak about is using source-associated mobile genetic elements to identify zoonotic extra-intestinal E. coli infections. And I'm sure by the end of the presentation, this will be more digestible than the title sounds like. Just a couple of housekeeping rules, please keep your microphone on mute. Please rename yourself with your organization and country followed by your name. Please note that Professor Lance's views are his own and not FAO's one. Please refrain from advertising your services, your company or any other commercial product or brand. And please post your questions in the chat, the box that you have in the Zoom. And the goal of this, it's called dialogue precisely because of that. We would like to have a lot of dialogue at the end and explore possibilities of collaborations between the speaker and the participants in this webinar. The meeting is being recorded, so please keep that in mind. And after it, we'll share the presentation and a couple of additional resources that Professor Lance shared with us. And at the end, we'll post a link for a feedback survey. So I'll stop here, Lance. And the floor is yours. Enjoy the webinar. Thank you for opening that up, Jordan. Hello, everybody. Good morning, good evening, good night. Let me open my slides. So you heard the official title. Here's my alternative title. Mama always said you could tell a lot about an E. coli isolate by its shoes. Where it's been and where it's going. And so the idea here is that by looking at accessory elements, particularly mobile genetic elements, we can start to predict what host an E. coli came from, regardless of where we isolated it. And so this is the official title. So, you know, I think it's important to just sort of start off with this fact that E. coli is a very versatile organism. You know, just looking at the broad host range, you know, we're you all are F. And so I think think about some of the major food animal species, at least in the United States, these are the dominant food animal species. E. coli colonizes all of them. It also colonizes our our interesting clip art person here. So it also colonizes us, right? So most most humans have some level of E. coli in their gut. And that E. coli is dynamic. And so, you know, from a from a molecular perspective, there are, you know, thousands of E. coli strains out there. You know, if you just look at the MLST database, you'll see, you know, literally thousands of different MLST sequence types. But from a clinical perspective, E. coli can really be bend, excuse me, into three different categories, you know, kind of the good or commensal, you know, benign E. coli that we all are probably carrying now. There's the dirogenic E. coli. Those kind of cause diarrhea, you know, like 0157H7 associated with feedlot cattle. And then there's the what I say super bad, but, you know, it's the it's the biggest killer are the ones that cause urinary tract infections. And those have lots of different monikers from UPEC urinary tract, you know, infecting E. coli to neonatal meningitis E. coli. But they all fall under this umbrella of extra intestinal pathogenic E. coli. So extra intestine. So it causes disease outside of the intestine. It's actually they're typically benign in the intestine. So they behave like a commensal otherwise. But when they get introduced into the urinary tract, they can cause infections. They get introduced into the blood, they can cause infections. And, you know, this is a big, big pathogen around the world. One of the biggest killers around the world. And and and the dominant organism when it comes to AMR deaths around the world. And so they have special virulence factors they can they can adhere to, you know, the the cells of the urethra and even under sheer force. Actually, they lock down even tighter under sheer force. They can cause bladder infections, kidney infections and sepsis due to or sorry, deaths due to sepsis. So in the United States, we estimate 36,000 to 40,000 deaths due to sepsis. So I just want to make sure that. Oh, sorry, I saw the I saw the chat came up. And I just want to make sure that wasn't for me. OK, I'm going to ignore it otherwise. So in, you know, we see that they're extensively resistant, E. coli strains becoming common around the world. The problem with this, you know, this extensive resistance is that if you can't stop an infection at the bladder, then it can it can ascend the ureters and get into the kidneys. And once it's in the kidneys, it has access to the blood. Right. And so, you know, when you present with a bladder infection, there's an opportunity. When you present with a kidney infection, there's an opportunity. But once it's in the blood, you have, you know, if somebody septic, you know, minutes count. And the more resistant and E. coli strain is the more dangerous the situation is. So I want to lock this relationship together just for those of you who don't realize. I mean, I learned late in my career that UTIs, most UTIs are caused by E. coli. So, you know, 80 percent plus of urinary tract infections are caused by E. coli. They can live in our guts without symptoms, as I mentioned before. And we get these urinary tract infections, abbreviated as UTIs here, when these strains make this short trip from the anus to the urethra. Now, women are a much greater risk for UTIs because of the anatomy makes that trip easier. But we've been asking this question, not how do people get UTIs? I think that that's pretty well established. But how do we get the UTI causing E. coli into our guts? Because, again, it's not all strains that can do this. And so we know that people can pass them, you know, pass strains from person to person. And this is almost certainly the dominant pathway, right? Is that we, you know, the world's covered in a thin layer of poo. And we're sharing E. coli strains all the time. And the people that we live with are the people that were most at risk for picking those up. And then, you know, sexual intimacy, et cetera, can increase risk. But we've also been asking this question, you know, can people pick these up from food animals? And this does not, this question does not originate with me. I mean, this question started with studies from the 1960s in the UK, where they had outbreaks of urinary tract infections in hospitals that were traced back to the suspected poultry products in the kitchen. And then people like Jim Johnson, Amy Mangus, Lee Riley, and others around the world have been studying this, the potential transmission. And really what we're looking at is, you know, not so much direct transmission from the animals, but probably via meat. Because when we look in grocery stores and we look at meat, it's, they're almost always contaminated, or sorry, I should say very frequently contaminated with E. coli. So several years ago, now we started a study in Flagstaff, Arizona, the small geographically isolated city near the Grand Canyon. There was no food animal production there. And we decided we would go to every grocery store in the city twice per month by every brand of chicken, turkey, and pork. We didn't buy beef because beef was not suspected as a source at the time. And we bought every brand, brought it back, cultured it for E. coli. And then we partnered with the only hospital in the town and got every E. coli from every urinary tract infection and blood infection there. And so by the end of the year, we'd cultured nearly 2000 E. coli from meat and nearly 1200 from UTIs and blood. Now, we also have, we have another study that we're just completing in California with nearly 7,000 isolates. So maybe next time I can present on that. So the big question here is, what's the population overlap between the E. coli in the retail meat supply and the E. coli causing infections in people? And this was sort of my first experience, not my first experience with naivety, but the first for this study. My big sort of naive thought was that I could just draw evolutionary trees like we'd done when we were investigating the Haitian cholera outbreak or when we were trying to understand the evolutionary history of SC-398 or other outbreak situations. I thought we could just use those models. And so our plan was to use a three-step approach, which was to sequence the genomes, perform MLSD. So we basically just extract an MLSD sequence type from the whole genome sequences. And then for every sequence type where there was both meat and human isolates, we would do a whole genome single nucleotide polymorphism based phylogeny. So in other words, a core genome phylogeny on each sequence type. So we can really understand those fine level relationships. So here's what the sequence types look like. So we found 456 different sequence types among the samples. Some more prevalent than others. And so the size of the bubble here, each number here represents the sequence type and the size of the bubble represents how many isolates had that sequence type. And if we isolated an isolate from meat, it's in red. And if it was from a human urine sample or blood, it's in this nice urine yellow here. This person clearly had some vitamins or is dehydrated. But so you have ST131, no big surprises, the most prevalent among humans. ST117 should not be a huge surprise either. This is a prevalent one, at least in the United States today in poultry production. And so we decided, all right, this is very interesting because none of them are very few, are pure gold or pure red. There's always this overlap. And so we started drawing trees. And we started with ST131 because it was so important. And most studies had suggested there was no food involvement. And in our own evolutionary studies suggested there was no food involvement. But I was intrigued by this big group of meat-associated isolates. So we performed a core genome phylogeny. This is an unrooted tree. Again, the same colors pulled up. So human isolates are in gold and meat isolates in red. And you see that the tree has these different clusters. So this cluster up here, H30, is the pandemic strain of ST131 that's killing a lot of people today. It's nearly, it's extensively resistant, fluoroquine, cephalosporin, etc. But you see all the meats are falling here in H22. Another FIM H, so this has to do with fimbri. But you also see human isolates clustering and some clustering very closely. So what we did was we just used the genomes from the H22s and we did a higher resolution tree. And this one, this time this is a rooted tree. And what we see is that we see some clusters, some clades where we have very closely related human and meat isolates, human and meat isolates. So we thought we had it here and we said, all right, this is clear evidence that there's transmission from poultry to people. But when we submitted the paper, the reviewers pointed out that, you know, hey, you didn't show us the direction and you don't really have a transmission and you don't really have clonal relationships here. These, there's a lot of variability here. And in fact, there's very few clonal relationships in this tree. And this really underscored the diversity of E. coli and one of our big struggles with this. And so we went back to the drawing board here and, you know, we started thinking about this diversity and the fact that we're using these tools that are really for outbreak investigations. And so, you know, think about, if you're a population biologist, if you're thinking about trying to define the E. coli that are colonizing, you know, 100,000 chickens in a poultry house, how many samples would you want? I'd want a lot of samples, right? But then if you scale it to, and I'm sorry for the US focus, but that's where I do my research. And so we'll just use the US as an example here. Most of our poultry production is in what we call the broiler belt, the southeastern United States, some in California, some up the West Coast, but mostly here. And each dot here represents a million broiler chickens. And so we produce 9 billion broiler chickens across the United States, mostly in here. And, you know, so now think about that one farm and you've got them scattered across here and how many samples would you want now? I think it's practically impossible to sample enough to really define the E. coli populations. And so, and we're using these tools are really best suited for outbreak investigations. You know, we're making trees and we're counting SNPs. That works when you have a young clone disseminating from a single point source, which is usually the case for an outbreak, right? So that's exactly what this slide says. So here's a visual for that. So you have a farm that has a new strain of salmonella, for instance, multi-drug resistant salmonella emanating from a single farm. Well, that's getting, you know, that's contaminating poultry products and then getting packaged up and disseminated across the country. But that will, you know, you can trace that back with, you know, product codes and looking at epidemic curves. You can potentially trace this back. Now, the industry's made it very hard to do that in the United States, but it is possible to at least trace it to a slaughterhouse. But the E. coli populations that are causing urinary tract infections, this XPEC, are old, diverse and diversified. That is that, you know, even around a single clone, like an SD, sorry, a single lineage, there's a lot of diversity, a lot of SNPs there. And so what you have instead of this, you know, single point source is that you have this constant bleeding over from the food supply into the human communities. And, you know, clearly I'm not drawing all the arrows here. This would be a very, very red map what would make a lot of Democrats nervous in the United States. So, but what I realized is that we really don't need to pinpoint a specific farm. What we're trying to ask, the question is not to, at least at this stage, we're not trying to say, hey, which farm are these coming from? Because we know it's, we suspect that it's bleeding across the entire industry. We just want to know, hey, which ones are coming from chickens or pigs or cattle and which ones are coming from people? And so that's a different, very different question and takes a different molecular approach. And so, you know, I'm a pretty simple guy and I thought, well, wouldn't it be great if E. coli just wore uniforms so we could tell where they've been, right? So when you see a person that works in a hospital in their hospital scrubs or their white coat, hopefully they're not wearing that in the grocery store. But when you see them wearing those in the grocery store, you know where they work. And I would love something like that for E. coli. And so this is my second force gump reference here. You know, so one of the scenes in the movie, I don't know if we're familiar with this movie, but it was kind of a culturally iconic film in the United States. He's sitting on the park bench and he sees this woman and he points at her shoes and he says, you know, those look like comfortable shoes. And then he says, mama, my mom says that you can always, you can tell a lot about a person by their shoes, where they've been, where they're going. And I can tell just by looking at this person that she's a nurse. She works in a hospital. She's got the white shoes, she's got the white gown. I mean, this is a nurse. And this is what we need for E. coli. And so I talked to my colleague, Tim Johnson, Timothy Johnson at University of Minnesota. And he's been studying these avian adaptive Col. V. plasmids for years. And he said, hey, have you screened your collection for these? And so we screened them. And sure enough, you know, 64, 63% of the poultry isolates had these Col. V. plasmids and only 5% of humans. And so from my perspective, you know, these really look like the shoes that E. coli wears when they're hanging out with chickens, right? And so this was our first hint at using mobile host associated mobile genetic elements as potential source trackers. And the cool thing about these is that they're unstable when you take them out of the avian host, right? So here is this L.B. broth. This is E. coli with Col. V. plasmids grown in L.B. broth. Those plasmids are stable. That's easy living for E. coli and L.B. broth. And here's chicken litter. So they're grown in chicken litter. That's really stable. Here's the chicken cecum. The plasmids are super stable. And this is a graph from Tim Johnson at University of Minnesota. This is what happens when he introduced the E. coli to the mouse gut. They drop that plasmid very quickly. And so this is, I think, pretty cool because over time, as they switch hosts, they're going to shed these plasmids and they'll probably take on new ones. And so we could use these to recognize recent spillover. And so the idea here is that we have a person currently smiling, but she gets a urinary tract infection so we can collect a urine sample and then culture E. coli. And if we see these Col. V. plasmids, that's a pretty good hint that maybe that E. coli came from chickens. And so we went back to our ST131 study and we decorated the tree with these plasmids, the Col. V. plasmids. And sure enough, what we found was that almost all of the poultry isolates carry these plasmids as did the human isolates that were most closely related. And so for us, this was pretty strong evidence that these people were picking up E. coli from poultry. Now, they didn't pick it up from this chicken, but they picked it up from chicken, the chicken population. And the reviewers were commenced. So we hypothesized that E. coli populations adapt to these different vertebrate hosts by shedding and acquiring different hosts associated mobile genetic elements. And Col. V. is clearly not the only host associated mobile genetic element. E. coli has a massive accessory genome. And so there's a lot of data to mine there. And so we started mining the data. And so we think that there's a lot of different shoes that they're wearing. And so by doing a large scale comparative genomics, we've identified 17 hosts associated mobile genetic elements in E. coli. And I'll just say that I don't think that we exhausted the full accessory genome here. And so we're continuing to mine. But we found 11 associated with one or more food animal species. And then we found six associated with humans. Now, all of our isolates were invasive isolates, or sorry, they were associated with symptomatic disease. And so we may be picking up here, you know, plasmids associated with virulence, but they were very strongly correlated. And so here's our heat map. Here are the strength of the association between these elements, H1 through 6, the human associated elements, with the human isolates versus meat isolates. And here are the meat associated mobile genetic elements association with humans versus meat. And you see obviously this inverse pattern. And what I'll tell you right here is that there's a lot of co-linearity among the different meat isolates. So chicken, turkey, and pork. And so we only use this in a dichotomous fashion, so meat versus humans. We're currently working on, by increasing our pool of genomes, by, as I mentioned, I think another 7,000 genomes, we're hoping that we're going to be able to break this down and be very specific about the different hosts in the future. So here's the next time I was very naive with this study. I thought we could just combine all of these assays together and that we would go to 100% specificity, sorry, sensitivity. But what I quickly realized is that as we combine these, we were really, we increased to over 90% sensitivity for detecting humans. These are the human assays, but our specificity started decreasing. So the more we used, the less specific they were. And it really got bad when we were looking at the meat isolates. So our specificity really plummeted down to between 30% and 40%. And so we needed a different model than to just combine these in a straight fashion. And so that's when I talked to my mathematician friends here, Dan Park, Jenga Wu, and through a medium, we spoke to Thomas Bayes, so the father of Bayesian statistics. And this was really the brainchild of Dan and Jenga using this Bayesian latent class model. And so I want to explain this approach using sort of a cartoon version here. So with the Bayesian approach, you have a person with a UTI, we get the culture, and we get an E. coli. Now we want to know did this come from a human or did it come from food animals or meat? And so initially we just have a probability, that E. coli, so it will come from a human or a food animal. But we do MLST and we get a specific sequence type. That might shift the probability. So now we have a different posterior probability that E. coli came from a particular host. And then we can start looking at our host elements. And as we run through there, that posterior probability will change. And you can start to estimate a probability that an E. coli came from a particular host. And so this is kind of a dumb explanation for how the model works, but this is the way I think of Bayesian latent class model working. The great part about this is that by training the model, then we could go back and run our entire data set and we could get a probability that an isolate came from a person or from meat. And this is really super straightforward. And so what we did was we said, all right, we had to go with 80% or better, and then we would call it from a particular source. And so that's what we did. And here are our urine isolates and the probability of a meat origin. So here we have our human isolates. Most of them have nearly 80%, have a 0% probability of coming from meat, but we have 8.4% of them that have an 80% or better probability of coming from meat based on the model. Now, this could be a sloppy model. So we looked at the meat isolates and we asked the same thing. And what you see is that almost none, 5% of the meat isolates had an 80% better probability of coming from humans. So this suggested, A, we handed all our samples well, that was a relief, and the model is pretty tight, that the model appears to be doing what it's supposed to. And we also obviously ran some formal testing. So we deemed these foodborne zoonotic E. coli and we estimate, again, about 8% of human UTIs in the United States, where we have very high quality water sanitation, hygiene, and food animals are produced far away from humans. We have about 8% overlap. Now, that might sound like a small number, but that exceeds any of the other major Europathogenic species like Enterococcus or Klebsiella. And when you translate that to a national level, we're talking about 480,000 to 640,000 urinary tract infections per year that may be foodborne zoonotic E. coli. It turns out that FSEC isolates are similar to non-FSEC isolates. So extra intestinal pathogenic E. coli that appears to come from people, they have similar, they're similar in terms of virulence. So here's the virulence data, the only place where they were a little bit different, sorry, they're the same when it comes to, not statistically different when it comes to cystitis, the place where they were different was, they were less likely to cause pylonephritis, but they were not different in terms of ability to cause sepsis. A few strains appeared to have an enhanced zoonotic potential. So here we have, we've plotted the prevalence of a specific sequence type among the meat isolates. So this sort of is a gauge of stochastic potential for an E. coli strain to sort of bleed over into the food supply. So the higher up it is, the more likely a human is going to come in contact with it by handling meat. And then the x-axis here represents the percent of clinical cases that, of a particular sequence type. So the further right on this axis, the more potentially virulent they are. Now this is, this is all cases, including asymptomatic bacteria. So that's just bacteria in the bladder. Here is symptomatic cases. So this is confirmed cystitis and pylonephritis and sepsis. And what you see is that a couple of sequence types really jump out, ST131, ST58, 10 and 69 are further right on this chart. ST117 is still right about 5%. And so it may be a little less benign but has a lot of stochastic potential. And then here are our sepsis isolates. And really ST58 jumped out. But this is a very small number of sepsis isolates. And so this is one case versus two. I'm not putting a lot of weight on this yet. We're doing some additional sepsis studies now. So when I think about this, and I talk about these high risk strains, and I talk about my, to my colleagues, like Tim Johnson at University of Minnesota, I find that some of these strains also cause disease in livestock, particularly poultry actually. And so by developing vaccines against these strains and applying them to the food animals, I think we could really have a win-win for public health and food animal production where we could decrease disease in the animals and then also decrease the bleeding over of these strains to people. And I'm really, I love win-wins. I often find myself, you know, angering the food animal industry in the United States by, you know, whining about antibiotic use and disease. But I think that this is one place where we could really team up. And when it comes to antibiotic resistance, you know, these strains were not, you know, they were fairly reflective of what we saw in the, among the meat isolates. But they were distinct from both meat and humans. And so these are our FSEC, so these are human clinical isolates that appear to be of zoonotic origin. You know, they were resistant to some important drugs, you know, and, but significantly less resistant than the human clinical isolates. And I really, you know, mark this up to the FDA's limiting of some of the most important antibiotics in food animal production in the United States over the past two decades. So we're, we still stand out among high-income countries in terms of quantity of antibiotics that we use, but we are mostly using tetracycline. And I say we, I'm not using any in animals, but the industry uses a lot still, but mostly tetracycline. And so I think that's prevented us from having the worst resistance patterns here. So these FSEC appear to be unique from both food-ass animal isolates and non-FSEC isolates in terms of resistance. But I think it's really important to point out, especially with this international group, that these strains could vary widely with geographic location because of differences in antimicrobial use. You know, it's a big planet and particularly the BRICS nations, Brazil, Russia, India, China, South Africa, we've seen this rapid, we're seeing this rapid development and this rapid sort of shift to meat-centric diets. And so we're seeing a lot of industrial, you know, industrialization of poultry production, pork production in some countries. And so this could be a problem. And we've seen, hopefully you've seen this Tom Van Boko study about antimicrobial, a previous study about antimicrobial use and this one about resistance rates. And what we see is the steady increase. And so, you know, I think we have to keep this in mind. And so likewise, I think the prevalence, just specifically the prevalence of FSEC infections could vary widely because of differences in environmental controls like water, sanitation and hygiene, but also the proximity of animal production to human communities and the level of subsistence farming. And so we have a paper that's under press right now about antimicrobial use and food animal production in Cambodia. You know, in Cambodia, you have over-the-counter use of antimicrobials in humans. You have over-the-counter use of antimicrobials in industrialized animal production. You even have antimicrobial use in subsistence farming. You have untreated human waste going into surface waters. You have animal waste going into surface waters. You have food being produced in those surface waters. You have open defecation. You have, you know, fish farms. And then you have some rudimentary slaughtering procedures. And so I think the ability for sort of bi-directional exchange is really underscored in an environment like this. And so I just want to wrap up with my last couple of slides so that we can have a lot of time to talk. And so, you know, our current efforts right now are to expand our host-associated element panel. And so we have, again, another seven or 8,000 E. coli genomes that we're analyzing, actually more than that because we've scraped the databases. We're looking to, again, get very specific with the host predictions. I don't know that we'll ever be able to split, you know, chicken from turkey. And globally, I don't know that that matters because turkey's kind of, you know, very regionally produced. It's popular here, especially at Thanksgiving. We want to improve host... That's what I just said, improve host predictions. We would love to develop an online tool with, you know, pathogen watch and other groups so that people can upload a genome from anywhere in the world and then get a prediction. We're also, and this is an important one, and I'm really putting a call out to you and to your colleagues, we would love to expand our isolic collections from humans and from, particularly from the dominant food animal species and low and middle income countries so that we can look at geographic variation and host elements and develop a geographically informed Bayesian latent class model that would allow people, again, to upload a genome or, you know, send an isolate for sequencing, upload that genome, say what geographic region they're from, and then get a prediction of source. And so if you have isolate contemporary isolic collections or you'd like to partner, please email me at lpricet at gw.edu. This is not a product. This is collaboration, George. Anyway, so that's the last of my official slides. Here's this, the requested slide here. And then I just want to acknowledge all my wonderful colleagues before I open up to questions. So my colleagues at George Washington University, Malia Aziz are bioinformaticist, Dan Park is an amazing statistician, epidemiologist, Northern Arizona University, particularly Paul Keim, University of Minnesota, Jim Johnson. Tim Johnson, Jim came up with the term XPEC. Staten Cerem Institute in Denmark, Mark Stegger, University of Michigan. So Jenka Wu and Mingbing Li, his doctoral student, were really pivotal in developing these Bayesian latent class models. And then, of course, my funders and my old colleagues at TGen. And with that, I'll stop my sharing and open it up to discussion. Thank you very much, Lance. And I knew it to be good. I didn't expect it to be this good. So it's brilliant that we aim to have these webinars on data-rich presentations, and yours could not be more data-rich. And yeah, regarding source attribution, and I think some of these data will be a lot of useful for risk managers later down the road. And what you mentioned, the connections with the different use volumes and the different antimicrobial use policies is something that is super interesting to further analyze. And you did it very well. The goal of these webinars is to trigger collaborations to find... We are all aware of the dimension of AMR, and we want this forum, this platform, to be used to find solutions, not just reinforce the problem. So yeah, like you said, it will be much appreciated if this will trigger sharing isolates with you from low and medium-income countries. And I see that you have a lot of participants from several different countries and regions. So yeah, I think you did it. Thank you for inviting them to collaborate with you. Thank you. Yeah, and we have some funds that we could pay for sequencing and shipping. And so I would really love to work with people. Brilliant. We got a question. It's an interesting one because you mentioned a lot about chicken and poultry. So you got a question regarding do vegetarians and vegans get less urinary tract infections? And how would this relate to your work? Yeah, so Amy Mangus is, I think, just a really brilliant epidemiologist. She left the field about 10 years ago, but prior to leaving, she'd done some studies looking at poultry consumption, poultry handling, and risk for UTIs. And there was a dose-response relationship. So the more exposure you had to poultry, the less, the more likely you were to get a urinary tract infection. And we know that the X-PEC are sort of predominate in poultry products. And so I think that that's pretty good evidence. I haven't seen great studies on vegetarianism yet. I think one of my concerns there is that I think it clearly could reduce your risk. If you're not handling these products in your kitchen, there's less risk for cross-contamination now, I would assume. But the problem is that the fecal waste from those animals is used as fertilizer. And so I think that you do have this environmental dissemination. And then if you're a vegetarian that also dines in kitchens where they handle meat, then there can be cross-contamination there as well. But yeah, I would suspect that it would reduce your risk. Great. Thank you, Lance. Next question, besides complimenting you, how do you think the host-level resolution of mobile genetic elements will work out, especially once low and middle income data is available, taking into account the leakiness between hosts and environment in the bi-directional, multidirectional transmission? Hey, I love that somebody used the term leaky. That's great. Leakiness. Yeah, we've tried to introduce this term. So that's brilliant. So with Maya and Adam Polly, who's now at Emory. So I think that that's going to be a challenge for us, is that as we go into low and middle income countries, I think the system becomes noisier. And what I'll just, this is between us in this room, we've been analyzing publicly available data. And what we see is that it appears that people living in low and middle income countries might have twice the risk of the high-end people living in high-income countries for invasive infection by these foodborne zoonotic strains. And so I do think that leakiness is going to play a role. Now the question is, how do we identify in a robust way mobile genetic, host associated mobile genetic elements in these leaky systems? Because the more they bleed together, and I think that's the point of the question, the more they bleed together, the less clear the signal is going to be. And so I think we've got to hold our judgment until we get in there and start to assess this. But we may have to develop our models as best we can in the higher income countries, in the cleaner setting, just sorry, less leaky setting, and then apply them in these settings so that we have that clear signal. We know what we're talking about in terms of host associated elements. And then, but it could be that there are industrialized animal production systems that are well enough isolated in some of these places where we start to get that. But again, it's an open question that we're very eager to explore. There you are. So maybe you've just started a new collaboration with Dan Sharpe, the colleague that's asking the question. The next one, what about the role of the environment? The colleague has done some limited work at surface waters and many of the nasty E. coli they found were expects and UTI associated sequence types. Could these faulty strains be environmental strains that are being shared between environment humans and animals? The one health perspective of this issue. Yeah, certainly. And I think that there have been some studies out of Japan looking at wild animals that are picking up some of the really important expect strains. Yeah, I mean, I think that there can be this movement of the strains from human populations into the environment through contaminated surface waters. There can be from the food animal populations themselves. Think about migratory birds. I think about what's called the Delmarva Peninsula here across the Chesapeake Bay from where I live. There's half a billion chickens that are raised out there. And then you have migratory birds that come in and they land and they feed around these chicken houses. And so I think the ability for these wild animals to pick up these strains and then carry them around is also big. And then the wild animals themselves might have their own avian pathogenic strains that can be shared with people. And so it's a world full of poo, right? But I think what's happening when you compare those sort of natural populations to what's happening in industrialized food animal production though, it's a very different risk formula because of the rampant use of antimicrobials. And so I think that that really elevates the risk because we talk about antibiotics being the foundation of modern medicine. But in some low and middle income countries, they're not the foundation of modern medicine. They are modern medicine. And if that antibiotic that somebody got out of clinic for a dollar or whatever, if that doesn't work, they could die because they can't afford anything else. And so I think that we really have to value these. And unfortunately this model, and I really want to own this for the Americans, right? We developed this industrialized model using antibiotics and we've exported it to the world as this sort of super efficient way to raise animals. Meanwhile, over the past 20 years, the FDA has started to ratchet back the antibiotics that can be used in a very sort of slow and frustrating way for me. But they have ratcheted back in a nuanced way. But that nuance, I think, has been lost for a lot of countries around the world. And so you're seeing sort of no holds barred use of antibiotics and animals. And I think that that's a huge threat. And then some of these, sorry, there's a long answer, but some of these countries, one of the ways that they're generating income is by mass producing animals and then exporting products. And so they can be exporting very contaminated, potentially dangerous products. Okay, thank you, Lance. The next one, again, more compliments, well-deserved compliments to your presentation. The colleagues have been recently looking at the potential transmission of E. coli from healthcare settings and community settings. There was high resistance for both, I believe, antibiotics resistance. However, they did not see much whole strain relationship but rather mobile genetic elements. Could this mean independent clonal transmission or how would these mobile genetic elements be moving around outside the whole strains? Yeah, so I want to make sure that we're clear when we're talking about mobile genetic elements. So I've got these host-associated mobile genetic elements that are not necessarily associated with antimicrobial resistance that I'm using to sort of identify host. But then we've got these mobile resistance elements, the R-plasmids, for instance, the resistance elements that are moving as well. And so I've been in my career, I've been mostly interested in the clonal transmission or the transmission of strains from food animals to people and really trying to quantify that. The other really important aspect is the movement of resistance elements. So you have the evolution of these new combinations of resistance elements in animals or in people that can then be transferred from one group of hosts to another, depending on wash conditions and other conditions. And I think the most striking example, kind of contrasting example that I've seen of this, and this is in the Maya non-empolis paper that we've, it's on bio-archives right now, but it's this leakiness paper that we wrote up, is that you see, if you compare the United Kingdom and the exchange of ESBL genes between food animals and people, there's very little overlap, right? So one of the wealthiest countries in the world, very high-quality water sanitation hygiene, separation of animals and people. Then you compare that to what we're seeing in Cambodia and there's tons of overlap with the ESBL genes. And the most striking thing that I saw there was that we saw this, we looked at the evolutionary tree of E. coli there. So we just looked at all the E. coli strains, we drew a tree, and then we looked at this one transposon that had the Q and R gene, so fluoroquinoline resistance and ESBL resistance on this transposon, and it was scattered across the tree. This really suggests that these things are moving prolifically in that setting, and we've never seen that here. I hope that answered your question. I think so. One question about the importance of the proper identification of the species, and the colleague is asking, what about the misidentification of some strains like Isirisi Alberti? Is this possible? Can this be happening? It certainly can be happening, but not in our study because we did whole genome sequencing and we identified the species by the whole genome sequence. And so I think in our case, it's clean and I think, yeah, it really depends on how you're doing your species ID, but I think that that's important. I don't know the virulence potential of that particular species. I'm not familiar with that anyway. All right. The next one, colleagues that have been looked at commensin ecoli from animals with a majority with MPC, BLA-EC-15. What do you think is so special about this BLA-EC-15 and not another bectolectomases? So that's the C-T-X-M-15, right? Yeah. So I think the big thing about C-T-X-M-15 is that it found true romance in the ST-131 lineage, right? So that was this perfect marriage for damage to humans, right? So ST-131 is this amazing colonizer. It's really sticky. Oh, sorry. I had it wrong. BLA-EC-15. Oh, you know, I actually can't speak to BLA-EC-15. I'm sorry. I was not familiar with that. No, sorry. All right. Good. The next one is a good proposal on the possibility of making a microbial committee that you can share information from time to time. Those that are interested in following up on your research, how they be able to, is there already a formal mechanism to change information about this or? Could you repeat the question? So the availability to make a microbial committee so that you can be sharing information from time to time? Well, you know, I'm happy to share any time you want to invite me back. I'm, and I think I would really love to develop an international network of people, particularly working on bloodstream infections, but also that we could, we could look at food animal species in the low and middle income countries. So I don't know the best way to, to establish that, but maybe George, maybe this is something that, that we can continue to talk about. Because yeah, I mean, I can tell you that we actively try to translate our work. So whenever we publish something new, we try to put it out in pretty plain English so that people can digest it. But, and we have a whole series of papers coming out. I'm still, I'm still embarrassed that I don't know much about BLAA EC-15, but I'm going to learn about that. Well, I can follow up on the afterwards the webinar. And the last question that we'll take just for the sake of time. A colleague is asking, why do you think people in Cambodia are not showing more endemic resistance illness, even the conditions presented in your slide, meaning increased exposure? Wait, could you start from the top? Why do you think people in Cambodia are not showing more AMR illness, given the conditions presented in your slide, meaning increased exposure? Yeah, I mean, I don't know the incidence rates in Cambodia, what I know are the, you know, all I know are the E. coli samples that we collected there, that, sorry, that Maya collected there. I can't really take any credit for the collection, just analysis, but just the level of resistance that we're seeing there. Yeah, I don't know who is tracking incidence rates, you know, globally as anybody. I mean, it's really, glass is supposed to start doing this, right? Yeah, from the human side. I don't know that we really have a good, sort of, across the world comparison yet. All right. Thank you, Lance. I think we will stop it here for the questions, and, but thank you very much for your presentation. What a privilege to all of us to be able to click on the link and listen to these presentations. And I think throughout the presentation and now in the discussion, several follow-up possibilities came up and a very concrete one, like you said, the possibility of colleagues to send you some isolates. And then from there, maybe you can reach the microbial committee that was also suggested in the chat. So I'll stop it here. Let's just share the upcoming, you can share, yes. So the next one will be on June 18. And this time we'll be focusing on Asia and the resistance to last resort across One Health in Thailand and neighboring countries by a professor Rungthip from the university from Thailand. And we'll send you out the feedback questioner. And if you have any further suggestions or ideas for this work, you can email us via the Antimocular Resistance at FAO.org. So thank you very much. And thank you very much. Thank you very much. Bye-bye.