Scientist Stories: Sean Conlan, The Human Microbiome

Hi, my name is Sean Conlon, and I'm an associate investigator in the Microbial Genomics Section at NHGRI. And today I'm going to be talking about the microbiome in general, and probably about half of my talk is going to be specifically on the human microbiome, which is sort of the more famous of the microbiomes. But hopefully I'll have convinced you by the end of this that there are lots of places where microbial communities are important. To life in general. So the microbiome is made up of lots of different organisms. And so what you can see here are some pictures, mostly micrographs of different organisms that you might think of as being in a microbiome. So in the upper right corner is a Staphylococcus. There's some fungi, there's mites, there's viruses. That's what that sort of large particle in the middle that actually probably a lot of people recognize now as a viral capsid. And so all of these things contribute to the microbiome, even though bacteria tend to be what we think of. And so in today's talk, I'm going to— in my introduction, I'm going to talk about the microbiome, the sort of main tool that we use to look at the microbiome, which is DNA sequencing, a little bit about the bioinformatics of how we analyze those sequencing data and visualize them. And then the second half of my talk is really going to focus on the human microbiome in health and in disease. And this is going to come from the perspective of the microbial genomics section that I'm part of, where our model system is really skin. So that's a little different from— if you're familiar with the microbiome literature, a lot of the microbiome literature is on the gut. I'll talk a little bit about that, but we really focus on the bacteria that live on the skin. So sort of the beginning of my introduction is sort of what is the microbiome and how do we study it? These are some sort of common microbiome sources that you might not think of when you think of the microbiome, but they are important. And so in the middle of the screen is a SCOBY. I forget what that stands for, symbiotic community of bacteria and yeasts, I think, which is how you make kombucha. There's yogurt, there's cheese. And these are all important microbial communities to sort of food manufacturing. And actually, as I'll describe in a moment, The cheese is actually a really excellent model system for understanding ecology and succession in microbes. And then I have the soap up there in the upper left-hand corner just to sort of remind me that in popular culture nowadays, it's really popular on TV and commercials to hear that soaps are gentle on your microbiome. I think that probably remains to be seen whether that's true or not, but it just goes to point out how much the microbiome has kind of entered the popular sort of mindset. And so these are some more examples of places where you might have seen microbiome in popular culture. This Gut Check game is a board game that you can play that you basically play microbes in the gut, and you're picking up plasmids and doing all kinds of different things in the gut. Ed Yong's book, I Contain Multitudes, is a very accessible book about the microbiome. And then there's lots of— sort of citizen scientist projects out there now because collecting microbes is a fairly accessible thing to do. But with all of this popularity and accessibility comes overhyping of the microbiome. And so what I have here are some examples, most of them taken from Jonathan Eisen's blog, on sort of overhyped microbiome. And there's probably a little bit of science behind all of these headlines. But in general, these are overinterpretations of the microbiome where some correlation, some correlated thing that was discovered in the microbiome correlates with a phenotype like athletic performance or something. And they haven't proven causation. They've simply shown that, for instance, with the bottom one about the cycling, that Prevotella may be important for— may be an important marker of athletic diets or the athletic gut. Like I said, could be, could be some science here, but for the most part, these are very overhyped interpretations of the microbiome. So what is the microbiome? And I already, I already touched on this earlier. It's bacteria, it's metazoans, it's fungi, it's viruses, and it's really anything that sort of lives in these communities. And I'm going to use sort of a general term of metagenomics, which is, you know, 10 years ago, 15 years ago, most of the time when you talked about sequencing something, you were sequencing an isolated organism. So you would take a bacteria and you would sequence its DNA, or you would sequence a human genome, or you would sequence a fungal genome. And metagenomics is the idea of sort of interrogating an entire sample to look at the entire composition of the sample. And it's very fundamental to how we look at microbiomes, how we take a census of what organisms are living there. And so these are some famous environmental microbiomes that you might not be as familiar with, but they really kickstarted a lot of what we know and a lot of the techniques and frameworks and models that we use to study microbiomes in a lot of other environments. So the first one on the left-hand side is an acid mine drainage, and this was made famous by Jill Banfield's lab. And the reason it's an interesting microbiome is that because acid mine drainage is so toxic and the pH is so acidic, very few things can grow there. So it actually makes an excellent simplified reduced model for understanding the microbiome. On the other side is the Global Ocean Survey, which Craig Vettner basically took a boat around and sampled places like the Saragossa Sea. And even though the microbiome is somewhat dilute in the ocean, they took liters and liters of organisms and spun them down or filtered them out and found a whole huge amount of diversity in the ocean. So everything from lots of different bacteria, cyanobacteria, and then viruses, phages, that are mostly associated with the bacteria. So these are two very famous examples of sort of environmental microbiomes that really shaped the way we look at microbiomes. I mentioned earlier that, you know, we all like cheese or Many of us like cheese. And part of what gives cheese its flavor and all of its, the thing, properties that we like about cheese is that there are microorganisms that are in there doing chemistry. We think of it as, may think of it as food science, but it really is chemistry. And it really is getting down into the physiology of what the bacteria are doing. It's actually an excellent model system for understanding those sorts of things, how bacteria are involved in the conversion of, from one molecule to another, but also in things like succession where you might, with cheese, you might start out with one community of microbes early in the cheesemaking process, and then that evolves over time and the composition of bacteria and fungi change over time. And that's actually not a bad analogy for what happens with humans, where the human microbiome develops from infancy to adults. And so Rachel Dutton's lab does some really nice work using cheese as a model system to understand sort of bacterial-fungal interactions. Another really interesting thing that you may not have heard of, but is a really cool system that Margaret McFall-Ngai's lab works on, is these little squids have light organs in them, which basically the thought is that when they're out hunting at night, these light organs glow and produce light below the squid, which basically erases its shadow. And so when they're hunting, they don't form, they don't create shadows. That's the, I'm not entirely sure if that is, if that's proven 100%, but that is the accepted reason for why they have these light organs. But they don't produce the light themselves actually. That light is produced by a bacteria called Vibrio fischeri.. And there's a really amazing cascade of biochemical processes that allow these squids to select one individual Vibrio fischeri out of the water column into this light organ, where it then grows up into a colony and is able to glow. And so this is, this is a very simple microbiome in some ways, because it's really, it's a monoculture of one organism that this squid has picked up to use for this purpose. And as far as the human microbiome goes, these are two very colorful pictures of ways in which people have tried to visualize the spatial orientation and sort of organization of bacteria. Taking the picture on the right first, this is sort of a look at the diversity, so how many different and how varied the communities are at different parts of your body. The redder it is, the more diversity and variation is. The bluer it is, the more sort of simple the communities are. And this is at the level of, you know, centimeter level on the skin. On the left-hand side are actually some very complicated micrographs of organisms from dental plaque. And so these are actually structures that are built up of bacteria in your mouth where they'll have— each bacteria is colored a different color using a probe. And so what you can see is all of these different amazing little structures where there'll be green bacteria on the inside with orange bacteria on the outside. Side, and those represent different species of bacteria that form these structures. And so this is a non-sequencing way to look at the structure and distribution of bacteria in these little structures. So I've already talked about it a little bit in the previous slides about how do we observe the microbiome? How do we figure out what's there? And a lot of it isn't using the microscope that's in this picture. I mean, that's really for sort of very simple, looking at sort of very simple slides and things like that. Most of the time what I'll be talking about is sequencing. And this is a slide that gets used a lot at NHGRI that sort of shows the cost for a genome as a function of time. So years are on the bottom. And all you really need to know is that the cost of genome sequencing has come down a lot over the years. And it's actually now very quite affordable to sequence human-sized genomes And while bacteria themselves individually have genomes that are, you know, many, many, many times smaller than the size of a human genome, in a metagenome or in a sample, there are many of those. And so we actually need to sequence them quite deeply. We need to sort of gather a lot of sequencing data in order to be able to really enumerate all of the different bacteria, fungi, and viruses that are in a sample. So the advances that have been made for sequencing human genomes have been a huge boon to the microbiome community. 'cause it really allows us to get in and really take apart these samples at the sequencing level. So how, nuts and bolts, how would this actually work? So in our lab, we might do something where we would call in either a healthy volunteer or we also work on atopic dermatitis as a disease. And so we might have a study where we're looking at somebody who has atopic dermatitis. Either way, we'd bring that person in, usually take a swab that's been moistened with some sort of a buffer. And use that to swab their skin. That's then put into another tube of buffer. We run a number of molecular biology protocols on that that essentially convert those long strands of DNA into small pieces of DNA that are tagged in such a way that the sequencer can see it. And then we sequence those in a massive way on a sequencing instrument, which I'll talk about in a minute. And then bioinformatically, we can do a lot of things with those sequences. So we'll convert a single sample from the skin into, say, you know, 50 million sequences. And then we can take those sequences, we can do some quality control and trimming on them. We can remove the human because part of what we get when we swab humans is that we get some of their cells. And so we'll end up sequencing part of their genome, and that's not part of what we're interested in. So we'll typically subtract that host DNA. And then we'll use things like public databases to take those, all the rest of the sequences, sequences and say, okay, well, where did this sequence come from? Did it come from a bacteria? Did it come from a fungi? Did it come from a virus? And so there's at least two ways that our lab looks at the diversity of sequences. We either will use amplicon sequencing or metagenomics, or what they call shotgun metagenomics, because you basically, with shotgun metagenomics, you're shearing the DNA up into lots of little pieces and kind of sequencing everything. Amplicon sequencing, which is what's shown on the left, is a way of looking at marker genes. And so every bacteria— so, and it's really a way to interpret— it's really a way to interrogate single kingdoms. So with amplicon sequencing, you can look at all of the bacteria in a sample, but not the fungi or the viruses. Or you can look at all the fungi, but not the bacteria and the viruses very easily. And viruses generally are not looked at using amplicon sequencing because they don't have a common marker sequence. But taking as an example bacteria, bacteria have a common marker that's called a 16S ribosomal RNA. It's very well conserved. It's, it's necessary and it has parts that are folded up and as part of the ribosome. And so they, they don't change very much evolutionarily. And the changes that they do pick up tend to give us markers for how bacteria are related to each other. So they allow us to make a phylogenetic tree. Or a tree of life. And so we can amplify those sequences using PCR, sequence those, and then use those essentially as tags or as barcodes to figure out what organisms were in there. It doesn't tell us as much because knowing, knowing that you had, you know, 5,000 sequences associated with Staphylococcus aureus does tell you a piece of information, but it doesn't necessarily tell you about what strains of Staph aureus in there. Or what plasmids they have, or what antibiotic resistance genes they have. It simply tells you that something from that bacterial taxa is there. Metagenomics, which is shown on the right, is a lot more comprehensive. It's more expensive, and the analysis is more difficult, but you get everything. So instead of just sequencing those little marker genes, we literally shear up the whole— all the DNA in the sample into small pieces. And sequence all of it. But what you get out of that is something like if you were to take multiple jigsaw puzzles and dump them together. We have lots of pieces and we have to sort those out computationally later if we want to try to put them back together as genomes. That's actually something that's a fairly big advancement in the field right now is that people are taking these heavily sheared up fragmented DNA samples and actually putting the bacterial genomes back together again to get what they call MAGs, or metagenome-assembled genomes. Most of what I'm going to be talking about today is not going the step of trying to reconstruct the genomes. We're just asking what genes are there, and based on those genes, what bacteria are there. So a little bit about DNA sequencing. A lot of people are probably already familiar with this. But DNA is a double helix. There are bases on each strand. There's a Watson strand and a Crick strand, or a sense and an antisense strand, and those encode the genes that are eventually transcribed into RNAs and are all the instructions for building a cell. And so reading that DNA sequence is a huge industry, and there have been huge advances over the last 20 years that have taken us from being able to sequence 96 samples, or 96 sequences on, on a run to being able to sequence billions and billions and billions of sequences in a run. And so something like the Illumina MiSeq DNA sequencer— and I'm not in any way associated with this company. This is just one of the companies that makes DNA sequencers. That instrument can take samples that have been prepared in a specific way and give you back all of the sequencing data. And without getting into all of the details, the way it works is it's basically your sequences that you're interested in knowing, we put little adapters on either end of those, which cause them to stick down to essentially a glass cell, or glass— think of it as a glass slide. And then we can watch the addition of bases to those as basically as we synthesize the other strand. And as those light— basically as those spots on a plate change colors, we can say that an A was added, a T was added, a C was added, a G was added. And as you watch that with a camera, you can then reconstruct the sequence of whatever, whatever was in that spot. And these are actually relatively short for DNA sequences, so they're on the order of, you know, anywhere from 50 to 200 bases, say. But they're very accurate. There are also technologies that allow us to sequence very, very, very long sequences at lower accuracy. And so the two of those technologies are the nanopore technology and which I won't have time to get into, but is essentially a way to take a small piece of a long strand of DNA and pull it through a membrane and essentially watch it as this sequence goes through a membrane, transmembrane pore, watching the voltage changes. And those voltage changes tell you about the sequences that are currently in the channel. And then there's also a technology that actually in a lot of ways is very similar to the way I described for the Illumina technology, where you're essentially watching bases being added, but it's at a single, molecule level. And that's the PacBio instrument that we use to sequence, for instance, bacterial genomes uses this technology. So in our toolkit for looking at metagenomes and genomes, we have these sort of two technologies, very short read sequencing where we can sequence highly accurate but short sequences massively in massive amounts. And then we have technology that allows us to read longer sequences but fewer of them and with lower accuracy. We see. And using these two technologies, we can learn a lot about microbiomes. This is an older slide, but it's just a very quick— give you an idea of if we were doing some sort of epidemiology, we were trying to find out about, say, a disease-causing organism, what are some of the methods that we can use to look at a disease-causing organism in a sample, and sort of how long would they take. And so things like PCR, where you may have a an assay for an organism, you might be able to do that in 2 hours versus sequencing the entire genome, going from the plate to the complete genome might take you 2 weeks. And so we apply these different technologies. When I get to the end of my talk about— we start talking a little bit more about sort of bad bugs. We apply these technologies as needed to get whatever sort of time and sequencing resolution we need because, you know, we get a lot more information from the right-hand side of the slide, but it takes us longer. We get less information, but we get it a lot quicker from the left-hand side. Other sort of omics technologies that people use to look at the microbiome include transcriptomics, which is looking at the, the actual RNAs that are transcribed from the genes. Because, you know, looking at just the genomics, looking at just what's in the genome, tells you what the bacteria can what genes the bacteria has, but it doesn't tell you whether they're being expressed or not, whether they're actually making that product. And transcriptomics helps you get at that. Culturomics is a kind of large-scale culturing of bacteria under different conditions. Proteomics is looking at the proteins in a cell or a sample. And metabolomics is looking at the small molecules. And these are all different complementary methods. Most of them still— there's still a genomic component to it, but then people will try to combine it with one of these other omic methods. So bioinformatically, you can imagine that when we do all of this sequencing, we get, you know, thousands, hundreds of thousands, millions of sequences. There's a lot of computational effort that's required to take that and convert it into essentially observations into sort of phenotypes and things you can follow up on. And a lot of microbiome work is what we call hypothesis generating. We're not always going in with a specific question in mind. Rather, we're going in to characterize a system that we're interested in. And a lot of times that characterization will create hypotheses that we can then go on and test in the lab. But one of the most sort of basic concepts that we are interested in when we look at sequencing data is the concept of diversity that I mentioned earlier. So on the, on the left-hand side is a low diversity environment, right? Those fish all look the same. It's all the same kind of fish. And there's lots of them. On the right-hand side, there's a sort of a coral reef and there's lots of different fish. And you can imagine that microbiomes can be the same way. You can either have microbiomes that are very low diversity. So for instance, your back or your forehead tends to be heavily dominated by a type of bacteria called Acuity Bacterium. Acnes. They used to be called Propionibacterium, but that's the dominant bacteria in that, in that community versus, say, the palm of your hand, which in addition to being kind of an unusual site, it's also a site that we talk about having a sort of high transient burden. So there are things on your hand from touching surfaces. So those tend to be sort of very high diversity. There's lots of different bacteria in different abundances. And we have all of these different estimators. An example that I have at the bottom that's actually not really used very much in real research because it's it's kind of an older estimator, but it's very easy to explain, is called the Chao 1 estimator, where you're basically just counting up the number of species you observe and adding to that a modifier that has to do with the number of species you've seen exactly once or the number of species you've seen twice. And those are— that modifier is in there to account for the fact that if you've only ever seen an organism once or twice, it may mean that you haven't sequenced deeply enough. If you have sequenced a sample so deeply that you've seen everything multiple times, that means you've probably sequenced it to completion. But if you're still seeing things only sequenced once, then that means there's still unidentified diversity in that sample. And this estimator helps you get at that. More commonly, we tend to have different ways of visualizing our microbiome data. So on the left-hand side is a fairly— almost cartoonish, but it is based on real data, image that we put out a long time ago that sort of shows what organisms live on different parts of the body. What you can immediately tell is that the oily sites like your forehead and the side of your nose and the back of your ear and your back, those are dominated by these kind of blue colors, which are Actinobacteria. And they're known to like these sort of oily sites. And so it's very easy to see it at a glance. And a more common way to see it is in the upper right-hand corner. You'll see these sort of bar charts where we kind of break it down and say that if a bar represents 100% of the bacteria, what percentage are associated with each different taxonomic group? And then on the bottom is another fairly common way to display lots of samples where we take the data that you would get from one of these bar charts essentially and compare you can compare two samples and say, okay, well, how similar are they together? And as you plot them on one of these, what's called— it's either an ordination, or in this case, it's a principal components analysis. Essentially, samples that have similar microbiomes are going to be next to each other in space. And so this, this one is colored by where the samples came from. And you can see that all the orange oral samples cluster together and all of the gastrointestinal samples cluster together.. And so it's not all that surprising that the microbiomes from these different skin environments form distinct clusters, but you can imagine in a more investigative sense, if you had samples that were, say, from pre and post treatment with antibiotics, you could see if, you know, after antibiotics, do the samples pre and post cluster together or do they separate out? And if they separate out, that tells you about that, what sort of a perturbation you're applying to the system. So this is kind of a summary of some of the different ways that we show microbiome data. And you'll see some more examples later on. So now I'm going to get to the human microbiome. And I'm going to start talking about the healthy human microbiome, or at least understanding it in terms of how the microbiome impacts different systems in the body. And so it's a fairly common trope that's thrown out there that there are 10 times as many bacterial cells in and on the human body as there are human cells in the human body, and that's because bacterial cells are much, much smaller. And that number has sort of been disputed, but it's not really disputed that the number of genes encoded by the bacteria that live in your gut, on your skin, and around your body encode many, many, many genes, thousands, tens of thousands of genes, and that these enzymes and these genes, these gene products are very important to human health. And just to give you a couple of examples, one of them is that some of the oligosaccharides that are in breast milk are actually not— there's not an enzyme in the human body that digests, that really digests them well, can break them up. But there are microbial— there are microbes that are able to break up these oligosaccharides. And it's thought that this is a symbiosis that's developed over the years, that essentially you're cross-feeding this, this microorganism. Metabolism. Also, the microbiome plays an important role in lots of drug metabolism, sometimes good, sometimes bad. Oftentimes we think of it in terms of the sort of unintended consequences. And one example I have here that's, I believe, from Dr. Balskus's lab, is that the drug that's used to treat Parkinson's is actually metabolized in the gut, and depending on the extent to which it's metabolized in the gut, that can actually impact the efficacy of the drug. So understanding how the microbiome is sort of doing chemistry for us, and sometimes in ways that we don't expect, is important. And it also contributes, it's been shown to contribute to lots of other different systems and disease processes within the body, including cancer, diabetes, obesity, and others. So as I mentioned earlier, the lab that I am part of, which is run by Dr. Julia Segre and our collaborator, Dr. Heidi Kong, we really study the microbiome in terms of the skin microbiome. And we have at least sort of two arms to our lab. One of them is, as I said, the skin microbiome, where we're really interested in sort of host genetics, the microbiome, and the immune system. And there's lots of diseases that we're that we're interested in that perturb one sort of leg of the stool. Atopic dermatitis, Wiscott-Aldrich syndrome, DOCK8 syndrome. So like, for instance, DOCK8 syndrome is an immune disorder that has a very strong effect on the viral microbiome. So by perturbing the immune system, we can now look at that effect of that on the microbiome. And then in the hospital epidemiology side of things, we're interested in sort of the genomics of antibiotic-resistant bacteria. And in both cases, we do studies with patient cohorts, but also model systems like mouse. This is a fairly famous sort of microbiome picture that came from the original Human Microbiome Project papers in 2012. And the main reason I have it up here is to show you this— what I was talking about is sort of what the bacteria do for you. And if you look at Panel A, where you see all of these different phyla, so these sort of large bacterial groups, so these are the, these are the highest level bacterial groups there are, you can see that there's quite a bit of diversity in how much blue or red or yellow there is. But if you look at B, at the metabolic pathways, these are instead of being colored by what bacteria they're coming from, they're just colored by what their function is. So what do they do? What are the what is the biochemical sort of repercussions of these pathways. And you can see that those are much, much more stable. And this is a huge, in my opinion, a huge conclusion from the Human Microbiome Project, is that, you know, the microbiome itself might be quite varied as far as composition, but what it's doing there tends to be very well conserved. In our lab, One of the more striking examples of a microbiome, of a change in microbiome, was this study from 2012. We looked at the skin microbiome as a function of age, and more specifically, Tanner stage. So if you're not familiar with that, Tanner stage staging is a measure of how far young men and young women are through puberty. And so Tanner stage 1 is basically toddlers or very, very young children, and Tanner stage 5 is basically adult. And there, a clinician can to essentially evaluate what tanner stage somebody is in. And you can see that there's a very clear demarcation in this top bar chart of the composition of the bacteria inside the nose. This is actually, that's where these samples came from, going from very young children, and you can see the ages on there, 2 years old, 3 years old, 5, 7, up through the older folks.. And this is such a dramatic change, and it's mirrored on the skin as well, where children are thought to sort of have very what they would call streppy skin. They have a lot more Streptococcus on their skin, and adults, or teenagers and adults, have more Propionibacterium or Cutibacterium. And there's a lot of reasons for this, but, you know, one of them is that as you age, your skin goes from goes to being more oily, and these oil-loving bacteria like Cutibacterium or Corynebacterium go up in abundance. And so this is a very clear example of the microbiome changing in response to human physiology. This is just another quick example of a study that we did looking at the interplay between fungi and bacteria on the human body. And so in this example, what you're looking at is essentially richness, which when I was showing you earlier that Chao1 estimator, this is a similar idea. So it's how diverse are the fungi or the viruses on the body. And what you can see is that for the plantar heel, the toe web, and the toenail, which are all on the foot, there's a fairly high fungal diversity but relatively low bacterial diversity. And that's reversed for the hands and the arms. Arms, and then the shades of kind of beige and pink, those are all on the trunk or the head. Those sort of have lower diversity overall. As I mentioned, one of the conditions that we study in lab is atopic dermatitis, and this is a dry, itchy, almost looks like a rash that typically manifests on the backs of the knees and the insides of the arms. For some people it's annoying, for other people it's actually quite serious and can pretty severely impact their quality of life. And the kids that we work with tend to be in the more severe range. So for them, it's very serious. And it also— atopic dermatitis is part of what they call the atopic march, which involves this condition oftentimes will then lead to things like hay fever and asthma. So it's an important condition to understand. And what we discovered early on is that that the disease severity, as basically when the disease gets more severe, the diversity of the bacterial community goes down and it actually becomes dominated by Staphylococcus, and very specifically Staphylococcus aureus in this case. And so what you can see is that we have controls on the far left, so those are kids that don't have atopic dermatitis, and while they do have some Staphylococcus shown in pink. You can see in the bottom, in the B portion, that most of them are kind of— it's this blue color, which is Staphylococcus epidermidis. But if you look at the atopic dermatitis patients, first of all, they show huge flares of Staph as they go through disease states, and that Staph is largely Staphylococcus aureus. So this is a big study that we did on really trying to understand the involvement of this one bacteria and how sort of like our discussion of the cheese microbiome earlier, how there is this succession of how the microbiome changes as a function of disease state. So it goes from this being a little more diverse state to being almost monocolonized with Staph aureus, and then going back to being more diverse after treatment is administered. This was all done with amplicon sequencing. This was done with metagenomics, and I'm just putting it up here to give you an example of We can get very similar information from metagenomics where we sequence everything in the sample. And in this study, we were really looking at, again, two important key skin species, Propionibacterium acnes and Staphylococcus epidermidis. And what we can see is that for P. acnes, that the P. acnes strains are very similar across a person. So within an individual, the strains that they have look similar between their the side of their nose, their palm, their toe web, versus Staphylococcus epidermidis, where actually the strains look more similar between different body sites. So everybody's alar crease looks more similar than it does to some other body site within that person. So these two different bacteria that both share the same niche, the skin essentially, have very different distributions sort of within or across people. So this is work that's not from our lab. It's from, from Eran Elinov's lab, and it's looking at the gut. But I think it's a fascinating study, and I think it's an interesting— it's something that everybody kind of knows about, and it's interesting to see that you can do this in lab. And this is sort of looking at yo-yo dieting. And without getting into all the details of it, if you look at the sort of top panel, the chart on the the right-hand side, that black line is what a sort of mouse getting normal chow looks like. And the blue line is what a mouse getting high-fat diet looks like. And weight is the readout. But the red line is what's really interesting, which is if you take a mouse and you put them on a high-fat diet, he obviously, you know, will gain weight as if he was on the high-fat diet. If you switch him back to normal chow or put him on a diet, basically the weight goes down. But then when you switch back to the high-fat diet again, instead of continuing at that sort of continuous rate,, he jumps back up to where he would have been if he'd just been, you know, eating the same high-fat diet the whole time. And this is the basis behind yo-yo dieting. Um, and so why is it that, you know, when you lose weight, you just gain it all right back and more? Um, but the really cool thing about this is that they can show in the, in the model system at the bottom is if they take a, uh, take a fecal microbiome transplant. So if they take the stool from a mouse that's either been exposed, um, to the normal chow or this, this high-fat diet, you can actually modulate where they sort of rebound to when you switch them to the high-fat diet, indicating that the microbiome actually almost has a memory of sort of where it needs to be, the sort of metabolic thermostat of where you're supposed to be. And so the microbiome has a role in this, in weight gain, obesity, yo-yo dieting, all that. So the last part of my talk is going to be on sort of bad bugs in the microbiome. And I don't typically like to talk about bacteria as being particularly bad or good. There is this sort of— what my boss calls— we use the sort of words of warfare when we're describing defeating bacteria or fighting bacteria. And in fact, most bacteria are actually doing good things. They're supposed to be there. But clearly, there are cases cases like this where our normal microbiome can be invaded by some sort of a pathogen, which can cause disease. And this is a very simplified model, and even as simple as it is, I don't know if it's exactly right. There is this idea in the microbiome world that microbiomes can become dysbiotic, where they don't look like they should, and that this makes them more vulnerable to invasion by pathogens. And one place where that's really— where people are really interested in that is this example from this paper by Tahra et al., where they were looking at stem cell transplants. And so I've blacked out the right half of the plot just to sort of focus you for a moment. And what you can see is that as a function of time going from left to right, this individual received a bunch of different treatments that are shown in brown. Those are conditioning treatments that are part of what they use to ablate the bone marrow so that they'll be able to receive the stem cell transplant. And so they received a lot of different conditioning regimens, and then they received a whole lot of antibiotics, which completely disrupts the microbiome. So you can see by the time you get to the middle of the slide there, the microbiome of this individual is almost completely this kind of shade of green, which is Enterobacteriaceae, which is Enterobacter, which is not the natural state of the microbiome. And so what they did was after the engraftment, which is when the stem cells take, they basically did an auto-FMT where they gave this person back a fecal microbiome transplant of their own banked stool, and you can see that it restores the microbiome. And the reason that this is important is that these, these heavily dysbiotic microbiomes, microbiomes that have essentially been disrupted by heavy antibiotic usage, are vulnerable to being invaded by something like Clostridium difficile, which is a huge problem in hospitals, causes a lot of problems. It's hard to get rid of. And actually fecal microbiome transplants are one of the things that has really kind of revolutionized the treatment of Clostridium difficile infections. And so this is a little fact sheet from the CDC on C. diff. If you go to hospitals, you'll tend to see these kinds of sheets on the doors. It's taken very seriously. But, and so this is an example of a clinical trial that was being used to to look at whether fecal microbiome transplants could prevent or treat recurrent diarrheal disease from Clostridium infections. And like I said, it works, but shortly thereafter you see articles like this. And in this case, two patients got very sick, one died from fecal microbiome transplant, and it was because of an antibiotic-resistant E. coli. Again, this isn't a hype sort of thing. This is more of a, you know, you have to be cautious how quickly you move on these sorts of things. You know, FMTs are, I think, going to show a lot of potential for treating certain things, but we have to be careful how quickly we go forward with it. And last but not least, these are— I'm going to show you just a couple of examples of individual microorganisms that we've looked at. So we have a collaboration that's between Julia, again, the Clinical Center epidemiologist Tara Palmer, and the Clinical Center Microbiology and Medicine group, Karen Frank. And we've looked at Klebsiella pneumoniae and a variety of other bacteria, primarily antibiotic-resistant bacteria. Specifically resistant to a class of drugs called carbapenems. And carbapenems are a powerful class of antibiotics that are used to treat fairly resistant infections, and resistance to this class of drugs has become pretty widespread across the world. And so we're interested in understanding the spread, especially because Carbapenem resistance tends to be on what are called plasmids, which are small pieces of DNA that can move between bacteria. Bacteria can very quickly become resistant if they acquire these plasmids. This is carbapenemase resistance in the United States. We've kind of zoomed into the US. It's a little bit of an older slide, but you can see pretty much everywhere in the United States has carbapenem resistance. These are examples of the drugs over on the left-hand side that they basically look like a little bit like ampicillin or penicillin, but they have all these sort of extra groups hanging off of them that make it a better antibiotic. But the mechanism of clearing these drugs by the bacteria is the same. They basically chemically cleave the ring where I've shown that red arrow. And so what we're always really interested in is this idea. So what I'm showing here is two individuals and a bacteria. That purple circle is a plasmid. That's kind of the classic thing we think of. We think of a bacteria maybe got transmitted between two people. But when you think of plasmids, plasmids are, like I said, can move around. And so we have to differentiate or we have to be able to detect situations like this where a plasmid may have gotten transmitted to a different host. And then that got transmitted or got transmitted to a surface where things like this happen. So for instance, a sink drain or a sink faucet. And we have to differentiate all of that from transmissions within the hospital from organisms that come in as patients come in. And so one of the big— from all of these studies that us and others have done, one of the big things is, you know, really screening and surveilling patients as they come into the hospital to understand what's in your hospital already so that you can detect transmissions and then work to prevent them. This is a slide that summarizes many years of work, but basically starting with a Klebsiella pneumoniae outbreak that we studied in 2011, looking at all the plasmids associated with different resistant organisms in the years that followed that, Looking at what these plasmas do long-term in patients that are colonized over many years, and then ending with a very large-scale study of what lives in the plumbing systems of hospitals. I'm just going to give you just a couple of quick vignettes that come from these studies to give you some examples of what we've learned by sequencing these bacteria. And it does come back to the microbiome idea too, because we're trying to understand what's living in communities in the human body as well as in the environment. So for instance, this is an example of two different stories from looking at bacteria and plasmids in the environment of a hospital. On the left-hand side, it's a fairly complicated slide, but basically two patients who never shared a room together. They were sequentially in a room. The patient that was there in November of 2010, there was a bacteria that associated with that patient, which resided in a sink, encountered a bacteria from a second patient who was there sometime later. They swapped plasmids, and in the end, what we observe is this sort of combinatorial organism at the bottom that has plasmids from both organisms. Organisms from both patients. And the complicated part about this is that what we originally observed was this sort of combined organism at the bottom that's labeled ECNAH2 that has 3 plasmids, all of which are carrying a carbapenemase gene. They're all carrying an antibiotic resistance gene. And we really had no idea when we first isolated it where all of these plasmids came from. And then we figured out, we sort of backtracked up a step and figured out this Patient A had the same plasmid. So we thought, oh, well, it must have come from that organism. But still, where did the other two plasmids come from? And only after we went back sort of into the freezer for other purposes did we find sort of the original organism. And so this is an example of a bacteria staying resident in a plumbing and basically in a drain on a— probably a biofilm for a long period of time. And I can say that once this was discovered, that the whole sink was basically pulled out and replaced. On the right-hand side of your slide is an example of just looking at how the plasmids are shared between patients and the environment. And the take-home message of this is there are a lot of plasmids that are only ever observed in patients or the environment, not both. There's actually not very much overlap in between there, which means that a lot of the organisms that are in the patient population are getting into the environment, and most of the plasmas that are in the environment are not making it back into patients. There's actually very few that span both environments. The last thing I'm going to talk about is something that we're getting really interested in lab. It's actually a fungus called Candida auris. It's popping up all over the world and then all over the United States. It has— it's It's what you're seeing here is sort of a phylogenetic breakdown of how they're related to each other using their genome sequences. And but the key take-home of this is that different clades of this particular fungus are being found all over the United States and Canada, Illinois, the sort of Northeast, Florida. And they belong to these different clades. There's a lot of work right now trying to track and understand the spread of this organism. And the reason for that is that Candida auris has some, has some really worrying properties. There are fungal species of Candida auris, or strains of Candida auris, that are resistant to all three major antifungal drugs, so they're very hard to treat. They can colonize skin asymptomatically, but then they can also enter the bloodstream and cause disease. And this is particularly true in cases of patients that have other comorbidities like diabetes or obesity or being bedridden. They can persist on environmental surfaces, including bed rails and windowsills. And they're, again, they're sort of, they're causing outbreaks in these areas that are very vulnerable populations. So these would be things like long-term care facilities and nursing homes. So this is something that we're actually studying actively right now, trying to understand it and trying to understand how the microbiome, you know, is there something about the microbiome that predisposes somebody to becoming colonized with Candida auris, either the fungal microbiome or the bacterial microbiome? And some of our work is beginning to bear out some, some, some conclusions about that. And then the last point here, it's listed as one of the world's 10 most feared fungi, even though it's only been really studied for the last few years. These are just some more notes on sort of popular, you know, both the medical and popular literature on this organism. So my final thoughts on this talk is that, you know, while it's the most famous, the human microbiome isn't the only microbiome. There's lots of environmental microbiomes. I didn't talk about it at all, but there's a huge literature on what's called the built microbiome, what lives in and on surfaces in homes, in homes and in offices, even in the space station. Over and over again, I can't stress enough that correlation is not causation. We tend to, when we're doing these experiments, look at a disease state or a phenotype or an observation, and we try to look at whether the microbiome is involved in that. And oftentimes we do see changes in the microbiome, but it takes, you know, making that observation is almost the easy part. Coming back in and figuring out if the microbiome is actually driving that change is much harder science. And that's where most people get tripped up when they overhype the microbiome, is not understanding correlation is not causation. And finally, microbial communities are complex, multifaceted systems. They're difficult to study, but we're starting to get a handle on it. We're starting to understand how we can intervene with things like from fecal microbiome transplants. We're starting to understand how organisms, how communities change as a function of other organisms come in using studies like the cheese microbiome where we're trying to understand how they change or how they respond to other organisms being put into the community. And how these communities are modified by what we eat, the drugs we take, antibiotics, all those sorts of things. And with that, I'll take any questions. Thank you. Okay, we good to go, Alvaro? All right, thanks, Sean. Yeah, that was awesome. Um, Belen just