 Hey folks, welcome back for another episode of Code Club. This is a very unique and different type of code club It's not going to be me coding. It's me giving a talk at the University of Michigan where I am a professor Every now and then we get asked to give a talk at our own department And that's what I did this past week on Thursday I think you'll like it if you're interested in science and kind of the tensions that we have between Reductionism versus wholism. I am much more on the wholism end of the spectrum And it relates to the research we do with the human microbiome There's not a lot of data analysis in it But at the same time like I said if you're interested in science and these types of debates that go on Among scientists, I think you'll get a lot out of it Let me know down below in the comments what you think and whether or not you think we need to be more Reductionist or more holistic. Thank you. So at the beginning of the year Kathy emailed me saying Pat It's been like 15 years since you gave a talk You're due and I was like a crap. What am I gonna say? and so I got this idea and I've been kind of workshopping it with Marcy over the past six months and when the U of M won against Ohio State and the National Championship like I got my title And then Kathy asks, who do you want to introduce you? And I'm always jealous of how our students introduce faculty when they have employment seminar speakers or each other for H12 So I thought I'll ask a student and so thank you for the introductions very kind of you So this is a brand new talk for me. I may never give it again. So We'll see so It's not as well practiced perhaps as the talk I've basically been getting constantly for the last 12 years So you're getting something fresh here So the team the team the team is a lift from a quote from the ocean back there Historic coach of the U of M who is kind of It when it comes to U of M and I'll share with you the quote later on But as as he mentioned in the introduction that I try to take an ecological approach to things when I came here You heard a few weeks ago from Joe Zacula that here's my first student and we actually came With with mason jars full of flour and flour beetles because we're going to study the microbiology of flour beetles And after going to our first departmental retreat that changed like instantly And so we've but regardless if it's a flour beetle or soil or the human gut We try to take an ecological approach Okay, and so as we go through hopefully the idea of the team the team the team will make sense and What you should know is obviously I studied the human microbiome and what I'd like to do today is share with you my critique in a way of human microbiome research and some of the external reactions that We hear as microbiome research folks And so one of the things that we frequently hear and many of you may hold this opinion That's fine. Is that microbiome research is too descriptive and it's too focused on associations And so we start seeing more and more of this chatter in more official things rather than just in hallways and so things like RFAs from National Cancer Institute saying that things that are not responsive to a particular call for applications Include things So it's a constant critique about against microbiome research that there's no mechanism We also see calls for papers and journals like Eli That's basically saying that we're calling for things that provide mechanistic insight into microbiome function, right? and Then on Twitter we see people quoting things that I've actually heard in this exact room at this exact hour that People prefer to do experiment omics rather than say metagenomics, right? And so this is a quote from the 2011 International Human Microbiome Congress or something like that so So maybe I have a bit of a chip on my shoulder And then there's this very old quote from Ernest Rutherford who won a Nobel Prize in physics many years ago That science is either physics or stamp collecting So stamp collecting is about the most damning thing you can say to somebody about their research, right? And I study stamp collecting. I developed tools for stamp collecting that I've gotten over 20,000 citations so I like stamp collecting and And so so what I'm going to talk about perhaps is a little bit of a defense of that and And and again critique also I think of how we've been performing microbiome research And I like to think in terms of analogies because although I look at data constantly and look at sequence data constantly It's very abstract even to me And so as a postdoc I had a wonderful mentor Joe Hammelsman who encouraged me to think outside the box And so this is a paper that we actually wrote looking at Microbiology if you will of various books using tools that we commonly use in microbial ecology To compare the community structure of those books and to see how they compare to each other And so you can imagine each word in the book representing a different organism or a different species of organisms and how often those words are used representing something about Their abundance right and I was the father of Some very small children and had read Good Night Moon one too many times and so that made it into this illustrious list of books and so The metaphor that I want to use for today's talk is this picture and so conveniently enough this is a community of about ten to the five individuals, right and That we kind of have a general sense of how this works right And so if American football means nothing to you don't worry I've only been in this building twice almost to get my son a COVID vaccine And I've never watched the event football game all the way through. Hopefully I don't lose tenure for that But if you look at this community How would we define health? Any ideas on how you might define health for a community of 110,000 people on November I think you can see it November 25th, 2023 So Saturday afternoon Thanksgiving season full Did we win Okay, those are great Since the Ohio State game So did we beat Ohio State, right? So there's some people that that's the only measure of health that matters for a given coach Did we end up winning a national championship? Did we end up going to a bowl game? What was our winning was our win percentage at the end of the season with us up with a with the seats full or not? Perhaps how many players were drafted into the NFL at the end of the season, right? So these are different measures of health and and this I'll tell you for a human microbiome is enormously challenging to define right because what we might define as health in one context is not healthy in another context, right? And remember that normal is perhaps not healthy, right? This might not be healthy, right? And So if we think about health and so think about it at this scale on this day of beating Ohio State There's 10 to the 5 individuals in there plus or minus Who is the most important individual in that community? JJ McCarthy. Thank you. Dr. Gunas By quam So these are quarterback and running back Does it do like me? Actually, no Thank you. Thank you. Yeah, not horrible Right, at least not directly because he's not here. He's sitting in the corner Doesn't matter that lights were still there. I won't I won't do that study, right? So You might you might say well, there's a hundred and ten thousand fans like that's a big Chunk of biomass that must be something in this community, right? Perhaps it's the the defense. I won't say host defense, but the defense that of Ohio State that just didn't bring their egging, right? Perhaps it's someone like Steven Ross or Uncle T or Connor Stallion's who all perhaps had some role in kind of the spy gate controversy that involved the whole season, right? and so again Thinking about who is the most important individual I think has a lot of parallels to this whole question of mechanism When it comes to the microbiome And so if we wanted to study this and figure out who the most important individual is How do you figure that out? Take one away so we could bench JJ McCarthy Right, we kind of did this with Harbaugh, right? Harbaugh was excluded He was knocked out of the genetically All right We could place the opposing team. I forget who they played before after this week. Does it matter? and Or we could have them play across the street at Pioneer High School and see if Having hundred ten thousand people mattered, right? We could go back in time to three thousand twenty and not let anybody in the stands, right? And so we could perhaps change the weather so But perhaps it could start snowing and we could see if that would affect that perturbation house, right? But ultimately what we could do is we could watch the game, right? So we could watch the game and we could see how the players interact with each other and how the fans interact with the players And perhaps you'd pretty quickly realize like maybe the fans aren't that important. They're not even touching the ball They're not in and around the players and perhaps there's energy, but how much does that really matter? and and we can perhaps So I guess we should step back and say like, you know, we know something about football and how it's played And so perhaps you could think about how we would not figure this out, right? So we we probably would not bench JJ McCarthy, right? We probably wouldn't do You know, we wouldn't probably replay the game in Chrysler Arena or at Pioneer High School We probably wouldn't focus on the abundant individuals, right? We probably wouldn't also do Koch's postulates, but this is a microbiology seminar. So here we go, right? And so to remind you of Koch's postulates the organism should be constantly present in animals Suffering from the disease and not found in healthy individuals. So of course, this is dealing with disease versus health And so I'm trying to talk about health. So let's maybe flip that, right? And so here we might think about You know, we're going to individual found with a phenotype. So again, we're going to leave out Harbaugh But we might think about JJ McCarthy or Blake Corum or Sharon Moore Right So the next thing we're going to do is that that organism must be cultured in pure culture away from the animal body And so we could do that, right? We could, you know, put them in the practice field or put them over at Pioneer High School all alone and Then when we put them back into a similar environment They should be able to do the same type of thing again, right? And then they should be re-isolated and and you know, we get the same phenotype back, right? So I'm thinking of perhaps someone like Danard Robinson who I think was playing when I first came to the OVM He was amazing and awesome, but the teams he was on were just never did much, right? And so perhaps he's someone that's really good in practice or in you know shots of light But in the whole context doesn't quite work, right? So my point here is that we wouldn't do something like Koch's postulates to show who the most important individual is in the vet community and So again, this is just way too complex And if you want to hear good debates on who's the most important individual in this picture, you know Any Monday morning in the fall you could probably turn on sports radio and get an earful and so my working model is That we have the team we have the fans and boosters and that if the team wins The fans and boosters will keep showing up and giving money to the team, right? And that the team is interacting with OSU And perhaps that's like a special game, but I don't as a fan. I perhaps don't really care about the coach I don't care about the individual players So if Sharon Moore next year wins a national championship Awesome, I don't really care who's the head coach, but that they won We obviously don't care about the controversy because everyone was very quick to defend Harbaugh here in Ann Arbor, right? We don't care about the type of offense they're running Right, whether it's a wishbone or West Coast offense or whatever, right? Just win And then we'll have a healthy relationship And I think this is replicated across the NCAA and so there's many many many football games going on any Saturday afternoon in the fall And they all have different coaches. They all have different players They have permutations on playing styles yet. There are some healthy and some not-so-healthy teams and So it's not about the specifics. It's perhaps more about the relationships and kind of the general guilds or Kind of roles that that people play and how they again interact with each other. So again Hold this in your mind. So again, this is too complicated. And so a Another sport that is near and dear to my heart And I think to this department is women's rowing You may not understand why but this is the 2016 women's eight At the Rio Olympics the women US women have won every Olympic gold between 2006 and 2016 They're just dominant, right? And so we might begin to think about This team and why these women Or the collection of women have been so dominant, right? So is it what defines their success? Is it the rowers the boat the fans? The training system that they undergo the water other teams, right? It's probably all of these many of these, right? It could be that you have awesome mentors who let you go off and train for the Olympics and win the gold medal And so that's Amanda Elmore who started working on a PhD and Graduated with gold and a masters sitting in the front front seat looking at us here And so Amanda was on a club team at Purdue came to the U of M with pips. I said sure why not go ahead train play with row with U of M She held the U.M. Erg Machine record as I understand it, which is basically one person rowing as fast and as Strong as possible, right? but they didn't win the national championship when she was at the U of M and and so she's you know as Dominant as she was She couldn't make up for all the other things to bring a national championship, right? So her on herself Wasn't sufficient, right? So again thinking about that kind of idea of coax postulates And and there's also considerable turnover in rowers that if you think about 2006 to 2016 There's a lot of turnover in the women that were in the boat and Amanda was only able to race in this one Olympics And so if sports aren't your gig you can think of other teams so to speak and so this is my team And that you might think about this configuration is being fairly idealized and by looking at the expressions on their faces You may think you know something about them and you may remain happy, right? But again, this is everyone being nice look at Papa for a picture and about you know One of about 10 pictures I took to get most people looking at me But then this might be them in a more natural System for themselves, right? And so again again think about microbiology and I could be the communities Do we study these communities and idealized situations? Or do we study them as they they actually live right and so this is this is a challenge, right? So in this picture, it might look like Simon Here is the only one doing any work, right? But you know, I can assure you that my wife Sarah is really telling everyone what to do here as usual Here's another team, right? And so these are the people that have worked in my lab over the past 15 years or so And I would be hard-pressed to say who is like the most important member of the team, right? Maybe it's the pig. I guess which you can't see over here But but really I'm the one that gets to stand up and kind of give the summary story and and Steal their good ideas to Get more ideas funded, right? And so I'm not this first person to kind of come up with this idea of thinking about Natural systems if you will to think about biology. There's a great paper from 2022 by Lasavnik from cancer cell that basically is kind of biologists fix a radio or what I learned while studying apoptosis Two things are trying to know nothing about I know what a radio is I have one of these I don't know if people still have radios or not try to explain what a walkman was to my kids and they're just like what But this is the inside of the radio, right? And so his idea was basically if you looked at this and you do nothing about radios and didn't have the schematics Could you figure out what the most important part of the radio was? So if we this is again getting silly and lots of metaphor, and I'm perhaps torturing this particular metaphor a lot And so this is my view of microbiology right now Is that we can kind of lose the forest for the trees and in some cases we lose the forest for a leaf or Stoma in a cell, right? and so It's again kind of like football That that we're kind of losing The the view of the football game for the spy plane Flying across which I didn't realize was like an inside joke that there was even a spy plane that flew over the game Right, so if you're just focused on the plane wondering what's going on there You're like missing the point of the picture, right? And so there's been kind of two general approaches that I've seen people taking To microbiome research, and I think microbiology in general So the first is what I'll call coax-based approaches, so we isolate and drill down To specific genes and perhaps codons even while working at the banter in some model system And so the mechanism that we discover through that context is specific perhaps, but not at a relevant level right, so this is kind of the the fable of Typically, it's the five blind men Feeling an elephant and describing different parts of it, but through the wonders of Google I was able to find one with lab coats Right, and so I would say this is important to understand the different parts of The elephant, but it still doesn't tell you really much about the elephant, right? And so still related to coax-postulates would be like studying Elephants but in a zoo right so oftentimes people focus on we've got a culture bacteria and to study their ecology And so again we can perhaps learn things about elephants, but I think perhaps one of the reasons why the Detroit Zoo Lost their elephants was because the elephants weren't doing well in a zoo, right? So we're kind of losing something in studying and trying to understand the elephants And so we can again come back another layer and study the ecology of these elephants And so taking much more of an ecological approach to studying biology and studying Elephants they in Africa and their natural ecosystem and seeing how they respond to predators This is lifted from an article talking about how animals sleep when there are things that want to kill them walking around Right, and so that's not something you're gonna get in a zoo where like we Protect little kids from knowing that there's predators, right? And so I like to think that this is more along the lines of what we're doing with microbiome research and and and and so I think that different sectors of society even from like scientists to Society in general kind of approach these three different layers As well, right? And so scientists love to reduce things and look at mechanisms of things And we talked about an early-life microbiota Metabolite that protects against obesity by regulating intestinal lipid metabolism, right? He was like, wow, that's it if I figure that if I like If I could get that metabolite, I will lose all this weight that I've gained from having all these kids Right and eating ice cream every night But again, it's a mice. It's with a particular organism the bacterium lactobacillus that lives in mice and is not particularly abundant and Juries still out whether or not it's actually doing this in humans, right? So we have a very specific knowledge of this system But how generalizable is it, right? But that's that's TBD Society also likes this kind of silver bullet mentality, right? And so when I was a kid, I recall my parents going on the grapefruit diet in the 80s to lose weight The Atkins diet in the 90s Fen Fen I remember as a grad student hearing about that getting pulled from the shelves Gastric bypass when I first got here and then Ozumpec, right? Which kind of wondering if I could get some so I don't I could keep eating my ice cream, right? But we want these like silver bullets That science tells people they can have and when I kind of look at myself And I think about my reasons for You know doubling and weight over the last 30 years. It's it's because of this picture, right? Like I hate wasting food. That's like ingrained into me as a kid. I have a fear of going hungry I think I'm gonna starve to death if I don't have seconds at dinner And I I love ice cream if you haven't gathered I love Big Macs and eating them fast I like eating right I get stressed when I get stressed I like to eat ice cream or chocolate, right? I don't like to exercise because I think it's boring. It hurts It's expensive to go to the gym, right? And so these are problems that are really important I don't mean to belittle them But it's not a single thing, right? Like none of the things that we're interested in is ever a single thing But I think when we approach things in science, we typically approach it as a single thing And so again, we come back to the idea of the team, right? That it's not a single thing There is no Bosch and Beckler would tell you there is no single player That's more important than the rest of the team, right? There's no one here. That's more important than everybody else, right? And and that's that's the truth, right? That we all have things that we contribute to the team For scholarship when we look at problems. There are many angles that we can look at that problem So I've served on stride through advance for the last seven years or so which is interested in figuring out how to improve representation in Higher by faculty searches, right? And it's been fascinating to look at people from across campus from the communities School of Music Theater Advance me Beth at one point Taking on this one problem and looking at it from different ways and kind of zoomed in and zoomed out And that's I think a very healthy way to approach things and that we really can't winnow things down to a single problem For any of the problems that we care about And so now I'm going to show some data And so one of the things to keep in mind When we think about studying the human microbiome is that there is no one human microbiome We are all different so this group that studies and finds this metabolite in a mouse Well whose microbiome does that look like right because everyone in this room has a different microbiome and what we see in this figure Is data taken from looking at I? Want to say 300 yeah 300 individuals who are highly healthy not normal, but healthy and They were then clustered into four different bins based on the type of bacteria in a fecal sample And the different font you know five of the most dominant or abundant most abundant Types of bacteria and we see wide variation right so like the back to 80s And some individuals is really high and then some individuals say like in community type B There's hardly any there yet. These people are healthy right by many metrics and We in this paper I'm telling you for a postdoc in my lab and I went through about 16 or 18 different body sites that were studied by the Human microbiome project we could then also look at the four different community types and see when these individuals were sampled multiple times How often did they move between different community types until we had fun that there are some communities like B? We're only 10% of the time that somebody stayed in B after first being seen in B, but other groups like AC and D were the most stable with D being the most stable right and so again Even within these three or four different community types As you can see by the size of these air bars and these box and whisker plots the rectangle is the 25th to 75th percent confidence interval right so there's wide variation in what is defined as normal And so when we study say a mouse or one person We're really looking at an idealized sense of that microbiome There you go Still blows me away 12 years later So this is work out of Jack Ravel's lab at the University of Maryland where they looked at the vaginal microbiome of women who? collected swabs over in this this woman's case I Think every day for 16 weeks or every couple days over 16 weeks And as this tetrahedron spins around you can see this red ball kind of bouncing around And each vertex on the tetrahedron as well as the center is one of five different community types that they found across many women that they looked at and Why or where the community moved they could not associate with anything right so they look at things like menstrual cycle Whether on hormonal contraceptives other practices They could find no association so sometimes it's that study I guess it already went through the loop But it these communities kind of like moved around sometimes they would stay sometimes they would move and I recall a physician at a conference I went to standing up and asking John like what why do you why is there so much variation? This this is abnormal. This is weird and jock is French. It's like it's ecology, right? It's ecology like these are things bacteria that are interacting with each other sorry for the bad front truck I've done work here with John the Puma's group who I think is now emeritus Although I still see him around so it's not totally retired And so his group has studied the cystic fibrosis microbiome and one of the frustrations of this project is that they would recruit You know a hundred people with cystic fibrosis and get sputum samples from them serially over time and there was no Consistent microbiome that these individuals had and if they're interested in like what causes a pulmonary exacerbation Every exacerbation is different right and so again, how do you do that? How do you study a system that is so? Variable right and again it would come down to each individual is their own experiment and and that's perhaps not satisfying to people that want Well contained problems, but that's that's the reality In my group, we have been interested in colon cancer and the colon cancer microbiome and so these are this is work done by Neil Baxter a former student at my lab and I hope you don't mind the gift but each line in the gift represents a different person who has a colonic lesion and the Represent the type of bacteria in their community and using machine learning algorithms We're able to distinguish diagnose people as either having a lesion or not even though there's wild variation among people who have Have lesions in their colons right and so there's signal there But it's it's far more complex than one individual one type of bacteria Right and so the other thing that comes to mind when we think about this is that the scale of our lens matters Right and we typically think that the finer the scale the better Right so in the microbiome field right now There's a lot of energy about like we should do everything at the amplicon sequence variant level or we should be sequencing metagenomes Which basically means that we're going to be looking at the finest possible DNA scale and And I don't think that's necessarily true or always warranted perhaps in some cases it is So a former postdoc in my lab Courtney armor looked at some of the data I just showed you and tried to analyze the community at different levels of taxonomic resolution from Amplicon sequence variant which we typically people try to think of like at the Species or strain level even back to the phylum level and then she built machine learning algorithms using random forest algorithm and Said how well do we do at classifying people as having a lesion or not? And so each each of the colored circles represents a different random number generator seed and that's black circle is the Medium plus or minus the 25 percent type and so what you can hopefully see Is that Amplicon sequence or ants? Is not the best approach right and it actually does about as well as looking at the family level which is fairly broad Right, and so we can also say peg the specificity at 90% and then if we change look at the sensitivity percent We again see that you know finer is not necessarily better that even as wide as like the family level So this would be like instead of looking at like E. Coli and Salmonella and such a vector Perhaps we're looking at in our bacteria. See it right. We're kind of zooming back taxonomically and so again If if everybody's colon cancer is different Then perhaps we need to look at a broader level to find those broader mechanisms that we're not going to find by drilling down to look More fine-scale We've done some work on Marcy Blunus's favorite set of metabolites the short-chain fatty acids and And we find much to many people chagrin no association between someone's disease status of colon cancer and Short-chain fatty acid right so the idea that you know there that if we find these silver bullets of metabolites then You know if we look at that find a scale But that will help resolve our problems and at least in our hands with our cohort One night I'm not able to find that there might be other metabolites and thankfully Marcy has an open mind and we're looking at those other metabolites as well so a case study that I want to share with you and thinking about kind of this this challenge between being specific and being broad or More focused and mechanism versus more general and mechanism I think is the question of obesity and This has been a Topic that has really captured a lot of people's imagination in the microbiome field and society And I think again it goes back to this the silver bullet mentality that if I fix my microbiome then I'll just shed the weight right And so a lot of this kind of got started and coming out of the Gordon lab Many years not that I'm almost 20 years ago now So Ruth lay had this paper looking at mice who were genetically predisposed to be obese So I think the OB gene is a leptin gene. You knock it out and the mice become obese And so they did heterozygous matings to control for the initial microbiome and what they found was that that if you looked at the communities in these mice they would cluster by by genotype and and that Well first they would cluster by the pedigree and then by their genotype and And that they also found a difference between the amount of formicities Versus bacteria that is that the obese mice had more formicities than bacteria that is as a ratio relative to the lean mice and so Again, this is what I'm talking about looking at a broader scale And so again, it's a nice and this question is like how does this play out in humans? And so because of course working with humans, I'll totally acknowledge is hard Praet Peter Turnbaugh then took The microbiome so sequel contents from lean genetically lean and genetically obese mice and then fed that into I'm sorry. Yeah, I Think actually this is from sorry this from humans so from obese individual obese an obese human and a lean person and then fed those into mice and They found that mice germ free mice who got the obese humans microbiota Had a greater increase in body fat When put on a Western diet relative to someone relative to mice that received a conventional diet and so again suggesting that this phenotype of obesity is Mediated through the microbiome and could be transferred to mice again. It was one human donor and so Who knows how much that scales and so we were interested in this question in humans and whether there was signal Through the microbiome of a person's obesity status. So basically if we've got fecal samples from people Sequenced them and then looked at who is in those communities. Could we distinguish people as being obese or not obese? And so some of these studies so each column across this slide is a different study So Baxter is from my lab Schubert is From my lab. Yeah, and then the other studies are studies. And so what we did was we said We looked across this many ways, but basically we trained a random forest algorithm To predict whether or not somebody was obese So in this first column we would take Neil's data and we would then say we know who's obese and not obese based on their BMI and We could then train that model to predict obesity status and then we would then apply it to the data from these other studies and What you can see is that none of the studies did particularly well at predicting obesity status in the other studies And when we look at things like diversity or this Bactroides and for Mickey these ratios None of it held up. Perhaps in individual studies it would hold up, but across the board There wasn't enough signal there, right? and so This gets me thinking or got us thinking that like I I believe the mouse data Right, but what are some of the reasons why the mouse data might not be translating to the human data and One thought might be just that there's tremendous variation beyond obesity status, right? And so Neil Sorry to put up the citation Neil Baxter when he was in my lab did a study where we collaborated with people at E.B. Memologists where they went up to I Have a daughter with Michigan Tech, so I can't call this northern Michigan. That's middle Michigan and so the middle Michigan to the biological station and Live-trapped parameskets which are field mice and then got stool samples from them every time they captured them And so there are mice that they captured multiple times from Two species of parameskets that overlapped in their habitat And so the idea of the microbiome is that like we have selected for our microbiome over the evolution And the microbiome we have is the microbiome we have and it's not like anything else and so what we found was surprising because Again looking across the top We saw no difference in metrics of alpha diversity between the two species. There's large variation, of course In what we found in those two species and in the bottom row are ecological distances saying how similar or different These communities are and that what we found was that with among these distances Within the species the distance or difference between the mice was pretty large And it's about the same size as the distance between species, right? So that would be like me and somebody else having radically different microbiomes This is this is actually much larger difference than what we find in humans And so that this this the distance is larger than what we'd expect based on these mice having been separated over ecological time and I thought this was kind of cool The reviewers didn't agree. I mean we got it. We got it published in applied environment microbiology But one of the reviewers comments has stuck with me since this was published like almost 10 years ago And and it's that's how much like reviewers comments can impact your work I think but this reviewer basically said that like well if you put them in a lab and Controlled everything then you could probably see a difference between these two species to which I said who cares right like I'm interested in how these these mice and their ecology their microbiomes are Are out in the wild where they actually live? They don't they don't live in a lab Basically what we're seeing is that all these other factors like the environment the diet their health is far more important than What species they're from and and I think that's Again coming back to this question of obesity is a question of scale And you might think of as like percent of the variation explained, right? So perhaps there is a role for the microbiome in obesity But I suspect compared to everything else the reason we don't see it in like our human studies Is that it's just a really small factor, right? So if you want to change something If I want to change something just stop eating ice cream, right? I should probably go run And not worry so much about my my microbiome and Again coming back to this idea of variation When I first got here and threw away the flower beetles and Vince gave me some of his mice to start my own breeding colony I had this question like well What does the mouse microbiome look like and so this is what assistant professor? I will just say a stupid assistant professor did I will say that about all assistant professors I was I went over to BSRB every day for a year actually two years and got a fecal sample out of a dozen mice So Christmas holidays all the time because I was interested in how the microbiome of these mice varied over time and They don't They don't and they're actually really similar to each other even mice in different different cages from the same breeding facility are Pretty homogeneous and so what we found in these blue lines is that in the first nine days post weaning There was a fair amount of variation, but there's a lot of instability in the community But when you get up to the five or so months that day after day And so this is kind of the difference in time between days and the difference between samples and the solid line is One mouse compared to itself over time and that what we found basically is like when we get out the mouse being old enough They they look like they're friends far more than they look A lot like they look like themselves right and so And again, this is in contrast to the perimiscus where these distances again Perimiscus by the same metric are like point nine Right and so our lab mice I think are good for a lot of things, but they're really homogeneous over time and it within the same breeding facility and so and this I think There are other questions. I think that come up with thinking about perimiscus But one of them being that a lot of the things we see in perimiscus We think should probably make a moss moss feel is really sick yet Whatever the perimiscus seemed fine, right? and So I want to quickly go through a case study here. I know I don't have a lot of time And talking about some C. The facility infection And kind of how some of these ideas relate to work in our own lab with see this and so Alex Schubert Who's a former student in my lab who now works at the FDA? Looked at about a couple hundred fecal samples that were obtained from the hospital as well as from the community Looking at C. The facility infection And so the the hospital samples were paired between people with diarrhea But no C. The people with diarrhea and C. The and then we also had non-diarrheal controls where people out in the community and so What you hopefully see here is that the gray dots in this ordination Are the non-diarrheal controls and they largely look different from people with diarrhea And but it's really hard to differentiate between people with and without See death that have diarrhea and and so You know we often wondered like well does a there's a diarrhea control really just a case waiting to get inoculated with C. Deaf, but we just we can't do that experiment, right? Which is why we do much And and there's other epidemiological factors that come up in these types of human studies so things like Antidepressants or proton pump inhibitors or a variety of other medications people are on that the epidemiologists tell us are associated with C. Deaf infection and so these are associations. We'd like to test them out from mice and So working with Vince Young's lab over the years We kind of jumped on their bandwagon of using mice as a model for C. difficile infections And so this is work from Nick Lesniak who followed up on a lot of stuff that Alex had done where we can treat mice with different antibiotics and we can then Colonize the mice with C. difficile after the community has been perturbed And we can look at the different types of bacteria that we think are most important to allowing C. Deaf to Colomize over time or to perhaps in like the case of kind of low Mycin and lower levels of self-preparation and streptomycin might even allow the community to push C. Deaf out of the community and Right and so we can then begin to think about testing some of these things or the ideas from the epidemiologists in our mouse model And so Sarah Tom give a transformer postdoc in the lab Was interested in well, what if we don't give them antibiotics, but perhaps give them peg We basically give them diarrhea, right? And so what she found in kind of this Probably overly complicated figure for this talk. Sorry. I know better is that We could colonize mice with C. Deaf without using antibiotics by giving the mice peg and Not only that like they would actually stay colonized Whereas treating them with clindamycin the mice could then clear C. Deaf a seal All right, so we can begin to think about other factors that Allow C. Deaf to colonize so maybe it's not just antibiotics, but it's some type of gut perturbation that allows C. Deaf to colonize We can also begin to think about drugs and affect other drugs might have as kind of Other factors that could help bring about C. Deaf infection. So one of these has been proton pump inhibitors The idea being that perhaps PPI's are changing the pH of the stomach and changing Kind of physiology of the gut and then allowing C. Deaf to colonize and actually what Sarah found in another study be published Was that a meprosol which is a proton pump inhibitor and she tried I think one or two others had no impact on sensitivity to C. Deaf infections, right and so One question about this that is like well is the literature wrong? Did we do the experiment wrong like why don't we see in mice what we see in humans, right? And so again, I go back to the the study that I did when I first came here as well as other work We've done and thought We're using mice from our breeding facility And so we're really looking at one person's experience with a proton pump inhibitor We're using a lot of mice to look at that person but We're not looking at a lot of variation in the composition of that community and So in another study that Sarah had done she bought mice from Four other four vendors and got mice from Vince's lab Which was the colony that we originally got our mice from and then ran the mice through the model again and what we find is that sure enough like if we get mice from different vendors they have different microbiota and When given clindamycin they clear C. Diff at different rates Right, and so it's probably not enough to look at mice from only one breeding facility that we need to build in diversity into our microbiome and So coming back to this Ordination one thought that we often had Was what if we could see the mouse with Many of these communities and run them through the C. Diff model And so Alex Schubert did this as part of her work and she took 15 donors Fecal samples and put them into germ-free mice and then tried to colonize them with C. Diff without using antibiotics. And so m6 Is the only door that had C. Diff so that's kind of our positive control and we used its strain of C. Diff Which is a more aggressive or more virulent C. Diff strain to then colonize these communities once they'd established in germ-free mice and The donors IDs with the red boxes were people who had what you might think of as normal or healthy microbiomes who were outpatient controls and so the M designation on the right side indicates those mice that died after one day of being exposed to C. Diff from the disease whereas the mice that donors on the left side Got colonized to varying degrees, but didn't die right and so again Same strain to the same genetics of mouse, but different communities got as different phenotypes So which community should we be studying? Probably all of them probably a lot more than these right, but it's hard and So this again comes back to the team that this stuff Is very complicated and one of my fears about the direction of microbiome research and science in general is that Instead of thinking about The game on November 25th in Ann Arbor We're beginning to play fantasy football where you kind of cherry-pick players from a whole bunch of teams and They don't have to interact with each other, but they're really good at whatever they do And that's how we're kind of studying things instead of trying to study things in the context that they came from And so I worry For my own field and every other person's field but mostly my own Because we're studying communities that when we when we extract one part of a community or team We risk making a scientist made fiction And so some of these things might look like would be like studying the event football team based on their spring scrimmage Right, like I'm sure it's fun. We get something out of it But like there's limitations to that, right? It might be like watching the football team play in Chris Chrysler or the basketball team play at the big house Right, like it doesn't really tell us a whole lot because it's a totally different ecosystem Different environment than what they're supposed to be playing in right? We can also look at unfortunately can look at oddball coincidences, right like I'm old enough to remember when I was at UMass That Michigan lost to this no-name school Appalachian State first game of the season when they're exposed to be like national champions right and so that's like a fluke and what people like you're going people hang your head at this perhaps but You have them still finished 18th in the country that year and still had a winning record and still won a bowl game, which was actually You know Jim Harbaugh didn't win a whole lot of bullies, right? So even though we have this kind of fluke event that's not indicative of like the overall season, right? We can also study people in isolation like Jayden McCarthy at the combine and I think there's a lot of people that go to the combine are awesome and then Lay an egg when they get to the NFL right and so we can begin to think about analogies for these in Microbiology and microbiome research and so my conclusions as if I haven't been opinionated yet enough yet Is that we need to be clear about the question and whether or not we can answer it, right? So I started kind of with a bad question like How is what is the health state of this picture, right? We probably want to be a little bit more focused than that, right? Perhaps you want to say like well What can we learn about football in general from this picture or lots of pictures like this, right? One game probably isn't that important I think it's really important for us to begin to think about like what percent of a phenotype Are we attributing to whatever mechanism we're claiming, right? Like we can find all sorts of mechanisms, but how important is that mechanism in the grand scheme of things? Also, the arrow doesn't always point in one direction right it goes in many directions that like the team winning makes the fans happy and brings more fans to the stadium which then You know perhaps brings more money in to name and likeness or whatever the boosters or like the nice locker rooms that the players have right It's a big circle But we get so focused that there's got to be a one direction thing and if you go to enough of these microbiome talks And I've trained people to say this we have a chicken or egg problem, right? Like what came first and the answer is like who cares they're both important You have to have both right to get more eggs to get to feed somebody And so I think just one direction mode is is overstated I think we should be setting bacteria and the host they're adapted to I say this is somebody that studies C diff and mice When like mice, I don't think it's C diff. So maybe we should study horses, but I don't think horse people let us do C diff and horses And I think we should be willing to embrace the complexity and context dependency of our problems and embrace the diversity of The systems and the organisms were studied right like we do a lot of work with say one strain of C diff Well, you know, we should do a bunch of different strains of C diff and how different see different strains of C diff respond in different types of communities But of course this gets things to be to be very complicated So finally want to acknowledge my team at least on campus and the people that have been in my lab over the years I'm able to stand here and do the stuff I do because of the contributions that all of these folks have made over the years as well as the many many many collaborators that I've had Since coming here and from before and so if people want to read it's actually really hard to read I don't I think it's probably better to hear him show that let's say it live than reading because it's a little bit Unintelligible, but the idea is that like when you're in college the team is what's important most important when you go to the pros But the team the team the team so happy to take any questions Thank you So the question was I've talked a lot about the embracing the complexity and diversity as their value in simplifying things and so I mean I definitely I think We need to simplify complexity to be able to say anything, but if I worry that sometimes we We focus too far down in on things We drill down on a problem and we kind of fail this Pull back and see kind of a broader context like so if we start moving variation like well, what are those sources of variation? I mean like for our mice our mice eat the same thing every day every meal. I don't eat the same thing every day for every meal, right? But We were controlling that sort of variation. So what effect does that variation in diet and just what you're eating? These are I think these are For everything we when we simplify things we need to understand what we're doing Right, so the question is Say like we study poop in humans and we can study other sites in mice like is there a disconnect there, right? Like does it does where we sample matter? I would say definitely And my lab is usually studied fecal material because we want to get a time course where we can kind of connect samples And so of course they're trade-offs And And Yeah, I mean it might be that if you look at a different place you see a different relationship, right? Like if we look in the cecum of mice or if we look at the small intestine, we'd see a different result, right? and so Yeah, we've done some of that like Caitlyn Flynn a former postdoc in the lab and also former graduate of our department worked with a gastroenterologist where we did unprepped colonoscopies on about 20 people to kind of compare the ecology off over the colon And so obviously like there's limits there like there's only so much you can say with 20 subjects But yeah, I think where you look is definitely gonna matter Yeah, so the question is basically like how much of this is a sociology of science problem And I would say yeah, like I think it goes back to that like physics or stamp collecting question Like like people want to be physicists They want to they want to nail something into the ground and say this is what it is We want the precision that a physicist has There was a guy who just I mean the frequency on microbiome meetings who run a sequencing facility and would say this isn't rocket science It's actually a lot harder than rocket science as we at least know where the moon is and we know we're like Cape and have all this but like when we do this we don't really have an idea of where we're going And it comes to training right like so I'm trained as an engineer as in the direction We don't think ecologically about things Maybe when we redo the curriculum, maybe we'll get more of that But I the other hat I wear is training people to do data science stuff People really are not comfortable thinking with complexity of data Or quantitatively, which is kind of required I had a department chair at UMass who scoffed at the need for statistics with the Added to I think was tongue-in-cheek like if you need statistics, you've done the experiment wrong Right, which I think I don't know that we see here, but there's messages of that across like biomedical research and That certainly doesn't foster the complexity that I know you all see and we see so I think if I'm getting it right so basically if we take the metaphor of Say instead of you event fans we gave all the tickets to Michigan State fans or Wayne State fans or different people Admission or just different people in general, or we perhaps we stopped out subbed out the team and brought in a different team And so in some regards we kind of did that with like the playoffs right like there were a lot of you event people there I suspect but it's totally different environment Perhaps we didn't change out the team Over time we perhaps change out the team, but bringing it back to microbiology You know perhaps something like that that just immediately comes to mind for some reason is like a fecal transplant And so people who have a microbiome that's been assaulted by antibiotics or some other perturbation And then now are getting susceptible to C-deaf that if we give them someone else's fecal microbiota they can then exclude C-deaf and not get recurrence and It's such an interesting effect It's like that because like people have tried to come up with cocktails of like a handful of buds or even one bug You can't exclude C-deaf unless you have like a whole community So that's kind of along those lines, but It's over time, but thank you all and I'm around if anyone ever wants to