Hello. Welcome to this webinar looking at the horse's gut microbiome. Metaboome and links to intestinal disease.

My name is Simon Daniels. I'm an associate professor of equine science and my area of specialism is, horse gut health, everything from nutrition, parasitology, gut microbiome. So this is an area that I'm particularly passionate about.

And the idea of this session is to Look at what we understand about the gut microbiome, the metabolome, and links to intestinal disease. So I've created this little graphic and it's quite busy, but it just gives us a bit of an overview because since. 2010 or even before we've really Advanced our knowledge of the gut microbiome.

We've known for a very long time, and I'll start with a little bit of history shortly, that the gut microbiome plays a role in energy production, synthesis of specific B group vitamins, for example, and we know that there is a role in intestinal disease in the horse associated with the gut microbiome. And over time, we've really started to realise that diets that contain high amounts of cereal grains with lots of starch are, are a risk factor for intestinal disease. Sometimes antibiotic treatments or other medical treatments can be linked to intestinal disease, as can various stressors.

Pathogen overgrowth and I suppose when we talk about intestinal disease, we could be talking about colic. We could be talking more on the digestive upset, really a sort of diarrhoea, really loose droppings, poor performance there, a variety of things, but with sort of colic being the sort of the end point that we're, trying to avoid. And so actually this tool has hugely advanced, all these tools that are out there have hugely advanced our knowledge and sort of along the bottom of this schematic here, there's the ongoing research, the different OMICS tools that are available which have really helped to improve our understanding.

With a long term goal that we can have better prevention if we understand a little bit more about the gut microbiome, the metabbonome, how these things work, it can help us to prevent disease. It helps us to promote better horse management, to prevent these things happening in the first place, but also, Sort of improvement, improving our approach to treatment or approaches to different therapies that might also reduce the, the risk of intestinal disease. And it sounds great and it is.

But we've learned a lot in the last 15 years, and as a scientist who's, this is my area, I am really passionate about this, but I also think we have to try and understand a little bit about what we now understand and what that might mean from a very practical perspective because It's great having additional data that we didn't have before, but actually, what does that mean for horse management? What do we really understand that we didn't understand before and how can we use this in a way that is useful to help us improve horse management? While also understanding that actually there are maybe lots of unknowns still and maybe this great tool isn't a silver bullet to fix lots of things, it's just slowly helping us unravel the black box that is the horse's hindgut, and I use this.

Term quite a lot this concept of the black box we know food goes in, faecal matter comes out, we get some energy, there's so much more going on in there, and every time we scratch the surface, we realise we're still just at the tip of the iceberg. I think going forward, it's always actually good to start by going back. And I've picked up a couple of quotes here, and we've got one from 1726 and one from 1897.

And really, this is just to highlight that for a very, very long time, we have been aware of the link between diet and intestinal disease. So the, the William Gibson quote from 1726, too much hay. Definitely excessive quantity of corn, may induce, indigestion and violent gripes, and new hay, and new oats combined are preferably known to be injurious.

So we've made some kind of diet change. We give too much starch, we see intestinal disease. It's still something we see today.

We're much more aware of it. We've known for a very long time that diet changes and high levels of cereal are likely to be problematic. Equally.

The the the quote from Ganji here from 1897. That grass is the first nourishment of all colts after they are weaned, whereas whether they're fed with corn and hay, but especially the first, so again that focus really on starch from cereal grains, then it exposes them to unspeakable injuries. So we've known for a very, very long time that there is a clear relationship between what we feed our horses and links to intestinal disease, yet we're still sort of unpicking what a lot of this really means.

And why a lot of these links occur. And the links to, cereal grains in a way, we, we know from an evolutionary perspective, the horse is designed to, ferment relatively low-quality fibre. So that doesn't seem particularly, surprising, but equally, we do use high energy cereal grains and performance horses and used in the right way, they can be really useful.

So it's trying to understand, I suppose, how to get the The best balance, and really, gut microbiome understanding, I think that the fairest place to start really is. The 1990s and the 2000s when there was lots of culture-based work going on. Labs such as Veronique Julian's lab in Dijon we were doing lots of culture-based work that really was our starting point for understanding a lot more about the gut microbiome, and a lot of this culture work was running in parallel.

With having, CCD cannulated horses, which again hugely understand our knowledge and advancement of what's happening with the, the gut microbiome, but there are huge limitations for the culture work. Because if you've got a culture-based study, you are extremely reliant on being able to grow the bacteria that you're finding in the, in the cecum or, further on in the colon. And many of those bacteria are strictly anaerobic.

We've, we don't have the right media to be able to grow them. So a lot of the earlier work was culture-based and could identify certain bacteria, certain microbiota that we identify in the horse's gut, but the limitations to that were there were lots of things that we could never culture and therefore just couldn't, identify them. But that definitely took us forward quite a large step.

The advances really started to come when we started to get the early, sequencing studies, the, the, the 16 using the 16S gene, looking at the bacterial DNA and this really kind of opened the door because the culture studies needed the bacteria to be live, to be viable, to grow, to be able to identify them. Whereas with the, the DNA-based studies, the bacteria could be dead. All we need to do.

I be able to extract that bacterial DNA and sequence it. We can then identify it against a database to be able to work out what is there. So this really was a sort of a great starting point, and I was fortunate enough that I was doing my PhD around the time when a lot of this early work was happening, and I remember my supervisor.

And I were having a chat in his office and he was fascinated by this new set of sequencing data, but it was full of unknowns and it was full of unknowns because at that point the databases weren't populated. We didn't know what the microbiota were. And if you were to take a lot of that sequence data today and reanalyze it against the databases now, we'd probably get slightly different findings because a lot of the unknown microbiota we would now be able to identify, but the early sequencing studies.

Using the sangu sequencing approach really started to open the door and really start to improve our knowledge of the gut microbiome. I mentioned these a second ago, and I don't think we should forget that we're actually indebted really to a lot of the early work where there were horses that were secretly cannulated, for example, and extremely rare to find now and for good reason, but actually at the time they've hugely advanced our knowledge of Function of microbiota, especially in the in the hindgut and the cecum, and being able to sample directly from those animals, and there are a variety of studies out there that look at various dietary interventions, giving elmintic various other treatment approaches, and the effect of that on the It could simply be the pH in the environment, but could actually be on bacterial counts again using that culture-based work and some of this then translates into some of these very sort of early sequencing studies that's really given us a good grasp on what's what the structure of the of the bacterial community profile of the horse is, which is really Helped to unpeel some of the layers of that, that unknown, that black box. Now, we often talk about sequencing studies, and, this is a sort of schematic slide that gives us a bit of an overview about what is actually happening.

So, 16S sequencing, there's so many papers now, looking at 16S sequencing, and it's, a gene, a gene that we can identify, in all, bacteria. And so what we have here is a, a workflow of exactly how one of these types of studies works. So you've got your bacteria, the bacteria have the, the 16S gene here, and we are looking just to amplify this, 16S gene, and we do this by PCR.

Now, the gene has different hypervariable regions. And so this is something to be aware of when you're looking. At different studies because some studies will use different hypervariable regions.

So then you have to question, well, are studies that are using different hypervariable regions, can you really compare the, the findings? Lots of studies use the V4 or the V3, V4 region, which would be relatively comparable, which are just different, or they are just a portion, of that 16S gene. So we gut the bacteria, we extract that bacterial DNA.

We amplify it and copy it up over and over again through PCR and then sequence it. A lot of the, the, so the systems like the aluminium MyE platform use a process called sequencing by synthesis, and they're barcode labelled, and you can identify all of your different samples, and you can then identify. The bacteria that are in your sample down as far as genus.

So you can start all the way up at the bottom and you can work your way down the, the taxonomic levels. Now, to be able to do that, you do need to also be a little bit of a computer scientist these days, or at least understand some, some basic coding. So at the end of your sequencing run, you have huge amounts of data, and effectively, the way that we deal with this is we work on, using either using a powerful server or you know, a specific computer set up in a terminal just like this.

Here we are, at the time we're recording this next generation sequencing, but the, the screens that we're working on effectively looked like some kind of 1980s computer setup. But you need to be able to understand basic coding to be able to do what's known as the bioinformatics, to be able to, use the data that's generated, and, apply that to an appropriate format to be able to apply the statistics, to be able to look at the, the composition, to be able to. Assign the taxonomy so we know exactly what we're looking at.

And if we're looking at two different groups, are there differences between those groups? How are those, how are those compositions. So there's the biology side of things, but there's also the, the sort of the computer science side of things.

And then there's an example here on the screen of a pipeline, CHM 2. And it's quite a common pipeline that's used for bioinformatics where you take the raw data and you do some cleanup, bring your, your paired data back together again, create a feature table, assign the taxonomy, look at the, the diversity, . And then you can do various statistics and plotting based upon that data.

So it is quite complex. And then because these types of data sets are complex in themselves, maybe with earlier studies you may have seen that certain approaches that probably wouldn't be overly appropriate. So quite often.

These data don't meet any form of normal distribution. A lot of the statistical packages that we would have thought appropriate in the past are often not appropriate. So there have been tools created and this here image here.

It's from a package. It's been around quite a long time now, but it's called LEFi, and this was designed to use multiple non-parametric approaches to be able to analyse the data. So that's all quite complicated, but effectively we want to be able to take a bit of horse poo in this case, extract the bacterial DNA, and actually get some useful data that we can robustly analyse to say if we've got.

Two groups, 2 groups often isn't very good, and I'll explain why in a bit, but if we create some kind of intervention study, so we have Horses that are untreated and then we do some kind of treatment, do we get a difference in the bacterial community profile and that bacterial difference or that profile difference is this. So that's kind of the, the technology and the tool. Now, amplicon sequencing is great.

And if we think back to the culture-based studies I was talking about, and how in the 90s we were really trying to sort of increase our understanding of the black box, we don't have to culture anything. Nothing needs to be live. So we can get the, the bacterial DNA.

It doesn't matter if it's dead, as long as we can extract the, the DNA, we can try and identify what it is. It's a relatively quick process and it's quite cost effective and in the time that we that this has been available, the cost has come down hugely. There are much newer technologies now, so as newer things come along and we find more effective ways of running assays like this, they become cheaper.

And they've certainly provided a lot of information about the bacterial community profile, the black box. It's, it's massively advanced our understanding. So there are some big benefits.

We have to also understand that there are some limitations, and we'll, I think we'll come and look at those shortly. So like everything, it's hugely advanced our knowledge on one side, but we have to kind of appreciate what we're, what we're looking at. What I have now are some examples of various studies, especially from some of the early work that really helped us to sort of characterise, our understanding of the horse's gut microbiome.

So what we have here are some, initial characterization studies. So the the first three studies here are all examples from sort of 2012, 2013. Kirsty Dougalls has picked out specifically on the side here, trying to understand.

The core bacterial community profile within the large intestine of the horse and really, the upshot of these three studies was trying to say, well, well, what, what is going on? What, what is present in the in the horse's hind gut? Is there a core?

What, what can we understand about it? Quite quickly, around the same time and going forward though, the, interest then starts to move away just from characterising what's there, but trying to understand. The bacterial community profile of the microbiome and various disease states.

So if we look at microbiome and horses that have colicked. Or we try and understand microbiome and metabolic conditions such as laminitis, or, the microbiome are, horses with suffering with equine grass sickness and comparing that to, pair controls. What happens to the microbiome, and I've picked out Joy Lang's paper from Joy from the, the University of Surrey horse microbiome Group.

Whose work the grass sickness work was looking specifically at what happens to horses with equine grass sickness to their microbiome, and so relatively quickly we started getting interested in looking at Different disease states and how this might well influence the microbiome, which again is a really useful tool and gives us a good feeling of how the bacterial community profile may well adapt or change in these different disease states. And this continued on, so, various studies have also looked, at antimicrobials, some of Mary Costa's work, identified here, looking at changes, in the microbiome associated with antimicrobials, which actually It's quite a logical expectation, but equally studies looking at proton pump inhibitors. So if you have horses that are being fed omeprazole for gastric ulcer treatment, does this have any effect on the gut microbiome because these just became unknown questions, .

But since that study was undertaken, the question I'd ask now is, does it have any effect on the fore gut microbiome, because most of these studies are focused very much on, on the hind gut, which is easy, or relatively easy to identify using faecal material. But actually, probably the bigger question would be, does it have any effect on, on the stomach microbiome, but that would be a completely different and slightly more tricky question to answer. Also, a group of us got quite focused on the effect of anthelmintic on, on the gut microbiome.

So work, from, a, a few different groups, in the UK, Ireland, and, in America, looking at the effects of parasites. Parasite treatment. So elmintic dosage on the gut microbiome, and I'm going to have a look at some of those shortly because these are all kind of focus areas where we can try and explore a little bit this relationship between, Things that are happening in the gut and the gut microbiome, interventions we might make, and what we might see, on the back of that.

So this all sounds really good and exciting and it is, and it has, really helped our understanding of what happens in the gut and in the gut microbiome. But I mentioned before, where we have, benefits to ampprion sequencing, there are definitely some, limitations that we have to, consider. One, and this isn't really just for amplicon sequencing, but as we've seen lots of amplicon sequencing studies, it was the, the first sort of technology that's become relatively easy for people to be able to access is there are different approaches.

So. You need to extract the bacterial DNA, and the simplest way to do this is to use a kit, and you could use faecal kits that are designed often not for equine but maybe for other faecal samples, human samples, for example. You can use soil-based kits which are often quite common as well.

But for a while it was quite unknown if the kit that you use may well influence the data that you get, and I had quite an interesting chat with a guy called Aaron Erickson when I was at a conference in the states earlier this year, and actually his group had done a little bit of work and found very little difference between the kits, which is reassuring to note, but it still for a long time was unknown. Equally. The pipelines.

So we talked about CIM 2, but there are various, bioinformatic pipelines that you can use and therefore, if you use a different pipeline on the same data set, would you get exactly the same results every time. And then when you align your sequences and look for the taxonomy, you label them against a database. Well, there are different databases, and so if you've got different databases and different pipelines and different kits, it sort of raises a bit of question as to how reproducible these studies are.

Now, generally when sequencing studies are published, all the sequences are there, they're freely accessible for people to go and get, and you could go and reanalyze them. So it, it, you know, this work can be done to try and really sort of get a better understanding of reproducibility, but I suppose my feeling for being boots on the ground is if you're looking at a study, that this is really interesting, but if you start comparing studies, you have to be quite careful in the method, you know, how was this and how was this approach undertaken and all these truly, comparable findings. The other thing that we have to think about is when we're looking at 16 studies, the 16S amplicon studies, they are.

Bacteria. You do get some methanogens, some archaea, which we don't see huge amounts of in the horse, they can be present in there, but we're not getting any fungi because fungi require the 18S gene to be amplified. So you're only looking at a small portion of the ecosystem.

And so that's just something to take into consideration. We've got very focused on the bacteria, but we could be missing a trick with the fungi. Then I've got a little schematic, and teaching anatomy and physiology at times, I think you always really need to think about anatomy or structure and physiology function together because standing alone anatomically, structurally, there's nothing wrong with a bike with square wheels.

I wouldn't want to ride a bike with square wheels because functionally, it's not going to be very functional. And a lot of the microbiome studies, especially the early ones, were very structural. What was present in that bacterial community profile, .

And it does look really exciting, especially if you do some kind of intervention and you see a change. Something's changed, but that doesn't necessarily mean that functionally there's a change. And even back in 2015, there was a lovely study that sort of was one of the first to show this approach that sometimes the bacteria can change, but the functional output stays the same, and we'll look at this shortly, but the term often bounded around it is functional redundancy.

Whereby one group of bacteria take up the job of the previous, so the output stays the same. And in a lot of the early studies, I think we got excited because we make this association between intestinal disease, bacterial shifts that oh, something's changed that could be our cause for intestinal disease. But actually you need to think about what the function is, what the, what the metabolic output is, because that really helps us to be able to identify what's happening.

Some of the other difficulties that we face along the way for steady interpretation is horse variation, and this figure, it's a few years old now. It's some work from Tina Blackmore, from the group at Aberystwyth, but this shows really nicely, and we see this often with both, amplicon sequencing and with metabolomics that we're going to talk about shortly, that. Quite often you'll get.

Variability between animals. So you get individual animal variability, and this can make studies challenging. So this study was a test train situation.

So there was a test diet, sorry, a training diet first, followed by a testing diet second, looking for differences. In the profile between those two diets. And what you can see here, there's 6 ponies.

Pony 6 over here, there's not a huge difference between the, the, the training diet and the testing diet, but the profile of pony 6 is quite different, the profile of pony 2. So there's In an ideal world, you would have a big cluster of test diets against a big cluster of trained diets, and that would show a difference in your bacterial community profile or metabolic profile between your two. Intervention or your intervention between the two.

But you often don't see that. You might see a difference, you know, pony 4 here shows some separation between the two diets. It's not huge, it's quite small, but that would be lost in the noise of the study.

And so that can often make the interpretation of some of these studies quite difficult. Especially because they can be quite expensive and though the cost is coming down significantly, the earlier studies would have been very expensive to run, so often the sample numbers were relatively small. So if you get a lot of noise in your data set, which is coming from that microbiome or metabollo, that actually makes it quite tricky to identify if you are seeing a difference after your intervention or not, which could be down to the individual animal response and the number of animals in the study, so.

This tool is fantastic, but it is not a clear cut approach. It can make it a little bit difficult when you're looking at the data and trying to evaluate how useful it is. Now we have moved forwards.

And it, it starts to become, it's even more exciting. So this technique here is the Oxford nanopore technique. I've actually got a, an image here.

This is sort of where 16S has moved forward to in the last few years. The amplicon sequencing approach that we looked at before, sort of looking at aluminium my seek is what's known as the little short read sequences, whereas long read sequences give us much better quality data. And the, the Oxford nanopore approach, especially using this little minion piece of kit here, has the potential to really take things forward in that one, it gives us long reads, which gives us much better quality data, and this example here, this little minion piece of kit is handheld benchtop.

Theoretically you could be literally out in the field. Definitely in a small practise laboratory rather than in a sort of some kind of genomics lab facility, being able to run these samples and it runs them literally in real time. So it's got these little tiny, motor protein nanopores, and you can see on the screen in real time as it's sequencing the data, completely portable, .

But it is still relatively computationally demanding. There was a relatively recent study out in the last year that looked, from some equine microbiome samples, being able to sequence the microbiome from skin. From saliva, faecal samples, there were some other body sites as well, and looking at the ability to be able to use this kit with the, the foresight that actually in a practise setting, you could have, this kit and you could relatively simply, undertake sequencing in-house and it is pretty cost-effective.

And that's a really exciting concept. There are times when this could be really useful. So if you were looking for a specific.

Bacterial DNA, for example, this could be a really good tool to have. I suppose scientifically, my question with the broader microbiome at the moment is I don't think we still know enough about what a normal microbiome looks like. So whether this is useful in veterinary practise in a practical setting as yet, yet it awaits to be remained.

I think in the future, I really hope there could be some exciting advancements that really make this useful, but for the time being. Certain, known bacteria, definitely, this could be really useful, but for broader microbiome studies, to me, the jury is still a little bit out. But this is definitely a way that we can improve on that, Amplicon sequencing approach.

Another approach, and this has been around a long time, is what we use is a whole genome or often referred to as shotgun sequencing. So what we have on the screen here, this is actually the 16s approach that we talked about. So with the 16S approach, you have your samples, so so this is your faecal sample, you extract the bacterial DNA and with 16S you amplify that 16S gene, and then.

Then you are using databases to be able to, they're grouped, the, the genes are grouped into similar groups, operational taxonomic units or amplicon sequence variants, they are now. You can identify what they are against the database, and then you can look at the composition and you can look at relative abundance, of what is present. That just gives you structure.

Shotgun sequencing, you extract the bacterial DNA, but instead of amplifying one gene, you amplify them all. So you looked at everything that's present in, in that community. Again, you need databases to reference them against to be able, and we can get a feeling for, to actually measure abundance of gene pathways.

So we get the function. As well as the structure because you still get the structure that you would get with, oh, I went a little bit too far. You still get the structure that you get with 16S sequencing.

So it gives us a lot more information. It gives us that functional information, so it gives us the, the, the structure and function together, which is fantastic. .

The only slight downside to this, last time I costed a project, we were still looking at 1000 pounds a sample. And again, like most things, costings are coming down, but it's really, really expensive. So it's a great tool, but a lot of the studies that have been undertaken are again on relatively small numbers simply due to the cost of being able to do this.

And again, going forwards, I'm sure the cost to some extent will still continue to come down. But we have to kind of look at what's out there and keep in mind that often this has been, the studies that are out there often on relatively small sample sizes sheerly due to the cost. There is some work again from, from Surrey, from the Horse microbiome Group.

It's been out a few years now from 2022 from Rachel Gilroy, but this was really one of the first studies looking at, 5 faecal samples and taking the shotgun sequencing approach. And what's really interesting here is it kind of took us back to where we started with amplicon sequencing. In that, what this study yielded, 123 high to medium quality metagenomes of bacteria in archaea, fantastic, but there were huge numbers of novel unnamed bacteria, genera, and down at species level.

Because we've never been able to identify them before, so they're not in the databases. So we're kind of back to where we started again at the tip of the iceberg. We're continually building and so we've got this fantastic new technology.

It's really exciting, that would be great to be able to get lots of money to be able to do a study taking this approach, but actually what you might end up with is lots of unknowns because we simply don't have the reference data yet to be able to fully understand what we're looking at. So if, if shotgun sequencing is really expensive and amplicon sequencing leaves us with structure without function, is there kind of a halfway house? Is there a way that we can try and understand a bit more about function?

Well, there are, there are approaches, and you'll find these again in various studies. So one approach, and this, is, from, Harvard from Morgan Langville's, Langvill's group, an algorithm that was the original one was created in 2014. It's since been updated.

The original one was called PieCrust. We now have PieCrust 2. And the idea of Piecrust is it takes the, the amplicon sequence of 16S data.

And it predicts what the metabolic pathway is likely to be based on the, the composition. And initially, it was 85% accurately predicting the, the metagenome. So yes, there's room for error, but this initially gave us a starting point for trying to predict if we have this bacterial community, we're likely seeing these metabolic pathways working.

This was then moved forward a little bit from an animal perspective and, hosted at Aberyst on a server at Aberystwyth, there was cow pie. Which was a ruminant version based upon pie crust. We now have pie crust 2 and pie crust 2 really sort of supersedes both these options here.

But pie crust will give us a, a, a, a better improved version now of predicting the, the metabolic, output. But it's still only a prediction. So another approach is to use metabolomics to look at the metabolites that are being produced, and I've got a little schematic here from faecal samples or urine samples.

You could do it from other. Biofluids, saliva has been using, blood plasma, but two non-invasives that can give us, especially gut microbiome are, both faeces and urine samples. And you take a sample and we, use it often use an untargeted approach.

We just look at all the metabolites that are present in there and then. To see if there are any specific group differences in a specific study design. And again, it uses, a sort of a computational statistical approach to be able to identify differences between groups.

And what my group have done a couple of times now is actually combine all three. So you take the Amplicon sequence data, this is the profile, you predict what that metabolic, pathway. Network looks like using pie crust and then you look at the metabolites that are present.

And if you have, an increase in bacteria that produce, lactic acid, and it's predicting that you're producing more lactate from pie crust and you're seeing an increase in lactate in the faeces, from the metabolomics, then you've got a fairly good indication that that is exactly what's happening. And that is one approach that is a more cost-effective way, to get an idea of what's happening in the microbiome. So I've got some various examples of studies, because I think it's interesting to kind of look and apply this.

So, this is a study from the, the, the last few years that, looked at longitudinal changes in faecal microbiota during hospitalisation horses with different types. Of colic. If we go back to 2010, some of those early studies were actually looking at the effect of colic on the gut microbiome and a little bit on different types of colic and also gave us some indication on some of the characterization.

The horses are not particularly good at mopping up lactate, they have, don't have huge amounts of, of microbiota that are good at converting lactate back, into propionate that you might find in our gut microbiome or or in a ruminant. This particular study looked at 3 different, colics in hospitalised horses, and quite nicely identified that actually the different types of colic classified. So in here there was an inflammatory, a simple obstruction and strangulating infarction.

You get differences in the bacterial community profile in different types of colic. Actually having the horses hospitalised for this period of time didn't influence the microbiome, but there were differences in the community profile and the survivors and the non-survivors in the inflammatory group. So there definitely is a role within the microbiome that can tell us a bit about disease.

We see disease associated changes and if we look back at some of those early studies again, we've seen before different types of colics show different bacterial community profiles and enjoying grass sickness. Study different bacterial community profiles in the horses with grass sickness, and we've seen it in modelled studies that are looking at metabolic differences such as laminitis. So maybe it's not a huge surprise that we see bacterial community profile changes that are associated with different disease states.

However, as I mentioned before, what can be quite tricky to then distinguish is this idea of functional redundancy. So Peppa Morrison's work here, and this particular study was looking at, the impacts of age and obesity on the gut microbiome, and they had a variety of factors. They're comparing, gross energy, intake and body condition score and various, markers, the glucose insulin, for example, in, groups of horses.

Against the bacterial community profile, but also looking at the bacterial community profile and the metabolic profile, the metabolo. And what was quite interesting was they identified a clear functional redundancy. So what you would see is a shift in the bacterial community profile, but the metabolites coming out were remaining the same.

And this is very much this idea of functional redundancy. So it's kind of like a backup system. So.

One of the bacteria gets swapped for another lot, but actually the job they do is the same, and it appears when this happens we don't seem to see any outward intestinal disease issues. It's only when we have a huge dysbiosis, a huge change in that bacterial community profile, that also changes the function. That we start to see those problems and this is really to sort of stress that we need to look at structure and function.

So again, when looking at these various studies, it's great to look at structure. It's taught us a huge amount, but we have to think about function at the same time. There are various ways we can think about function.

This is some work that I did, we published six years ago now, looking at the effects of anthelmintic on the gut microbiome. And the rationale for this goes back to the late 90s, early 2000s, again, looking at epidemiological colic studies, looking at risk factors for colic. There are so many things that are risk factors.

We know and have known again since the 90s that no parasite control. Is a risk factor for colic, but equally, a couple of, studies in the late 90s, early 2000s, from Noah Conan and others and from Mark Leary and others identified that anormintic treatment can be a risk factor for colic and specifically, with Mark Hillier's group, they identified that macrocyclic lactones, were, a risk factor for colic. So we wanted to explore this a little bit.

So we took a group of horses that had very low parasite burdens from repeated faecal egg counts, and they were already going to be treated. So we just observed it and sort of intervened. And we looked at the effect of their routine winter moxidectin treatment, on the bacterial community profile on the metabolome, and we also did some functional fermentation in a lab-based setting.

And what we identified was when we gave, half the group, . Moxodectin and we kept half the group as a control, but we also did it as an intervention study so we could compare that to look for the, the individual animal variation. In the fermentation study, what we managed to, identify was.

There was no difference in fermentation profiles of hay before any moxidectin was given. After moxodectin, it reduced the ability for the microbiota in the system to ferment the hay as effectively as it had pre-dosing, but very quickly this returned back to the baseline level. So within, 40 to 48 hours, actually, it returned back to, to its baseline level.

The bacterial community profile saw very little change, but one week after dosing the horses that had moxodectin, we saw a small increase in diffrobacter, which we associated with potentially the insisted saphostoins starting to emerge out into the gut. Which could be linked or have been linked specifically to this specific microbiota. So we only saw a very small change.

The upshot of this is the effect of anthelmintic itself is really unlikely to have a large effect on the gut microbiome. But at a similar time. Different group.

This was Laura Peachey's work. Laura looked at, horses with higher, parasite burdens. So she had groups that had very low faecal egg count, associated parasite burdens and higher where we would start to intervene with anklemtic treatment.

So for a group of young stock as well, so definitely we'd be more likely. To intervene and what Laura clearly identified is actually having a sciatostoin burden in the hindgut had a greater effect. These horses were also treated with with ivermectin, but the area that I've focused on here is where you've got your your different treatment groups.

The high group have a different profile. To the low faecal egg count group and the low group over time overlap a lot more, whereas the group with a higher burden over time we're seeing this dispersal and differences in that bacterial community profile. So kind of bringing these two studies together, it's likely that, if there is any link between Parasites and the gut microbiome, which there will be because they are, parasites are really macrobiota.

They're they're sort of interwoven in the gut microbiome. It's potentially, any risk of colic is likely to be a combination of parasite presence and death. And, and the reality is we still don't know the answer to what that link truly is.

Another consideration is really thinking about actually gut microbiome stability, so is it stable over time, what happens over a year? And really nice study here again, this is a few years old now, but this is from Chevel Salem. And others did a really simple study.

They had a group of mares out in the field, and they sampled them repeatedly over 12 months. And what they identified were bacterial community profile changes over time, especially in the late winter early spring when these horses ran out of grass. They were being supplemented with additional forage in this case.

Hailage led to a drive in that bacterial community profile shift. And I suppose for me with my nutrition hat on, I think, well, the grass profile, nutrient profile, maturity changes over the year, which likely also drives some of these changes. So you can't expect it to always be the same.

And we recently did a couple of studies. This is just a snippet of an abstract, and it was recently presented the European Workshop on equine nutrition in last summer, whereby we took two samples as part of another study, but beforehand and looked at, at, you know, did you get exactly the same bacterial community profile, same sample, same horses 24 hours apart. No, the difference is very small, but it's subtle.

If you look at the main two phyla, bacteriodda androbacter, we actually see a slight differences over a 24 hour period because the fluctuation and you get fluctuation in the faecal samples that are being, produced. So there will Always be a little bit of animal again variation, so we can't concrete say that a change is always 100% down to some kind of intervention. Very subtle changes could actually just be microbiome fluctuations or they might not.

And so these are just considerations that we have to have when we look at our different studies. So we've recently published a couple of studies that are just out. We did a nice study where we fed a group of ponies 4 different haages, and we wanted to look at the functional effects.

So this was looking at the metabolome and looking at the metabolic profile. So again, those studies previously that have looked at what are the risk factors for colic. And one of which is a change in forage, and hailage is really difficult because you can't change haliage really slowly over two weeks because it would go off.

So we did a study whereby we used a 3 day changeover and first of all, we did a characterization study. So we looked at the metabolic profile, the metabolic phenotype using metabonomics from faeces and urine of this group of ponies. And what we identified from 4 diets, our rye 1 was Very high energy, very low fibre, high water soluble carbohydrate was very different from the other three hailages.

The second rye hailage was much later cut. We had an early and late cut meadow, and we did a 3 day changeover between these different hailages. And what was really stark was the rye 1 was very different from all the others.

You could change over 3 days and definitely the . The three halas are very similar. The the metabolic profile between those halas and the changeover was quite straightforward, whereas the rye one was very different, and we got some marked differences in the metabolic phenotype.

And this really tells us that actually again it's useful we learned things we learned that actually at the very young Hailage we could clearly see. Metabolites associated with protein digestibility that would have happened in the small intestine. So it was a really useful study to give us an idea of markers as to what's happening in digestibility within the horse, but at the same time gave us an idea of what happens in this changeover between forages.

It'd still be cautious with hay. It always change slowly over 14 days if possible between completely different batches and especially where they're very different. And with hailages, ideally give a 3 day change.

Over between them, it showed that the gut's quite robust. The horse's hindgut's quite robust when the haages are very similar in their chemical composition. But when you get a hailage with a very different nutrient profile, it does behave differently in the hind gut, and we don't know if there's necessarily a risk of colic associated with that, but we do know there's a risk of colic associated with rapid changeover of forage.

So taken together, it would seem sensible to err on the side of a little bit of caution. Another couple of studies just to kind of draw this together. So this is a different approach, but this is a study we published last year.

I showed you a previous study from my group where we looked at the effect of giving elmintic moxidectin on the function and the bacterial community profile and metabonome. Well, this little study was a lovely example of using bacterial community profile, functional fermentation, and the metabonome. Looking at what happens in this little system, which we can use in the lab to, to effectively act like a little horse, hindgut.

And so over time, we took replicate samples from these little horse hindguts and while they were fermenting feed and we could look at the, the practical fermentation, but we could also look to see what happened to the bacterial community. Profile and the metabolic profile. What we see actually when we set this system up, we see some changes from the faeces to the system, and over time it starts to get closer to the faecal sample that it started with.

And again, we see this in the metabolic profile. I think this is a really nice example of how you can take 16s amplicon sequencing, which this study was look at the predicted metabolic profile, look at the metabolites and bring them all together to get a really good idea of how structure and function are working together. And I think sort of on the, on the study side of things, finally, this is a really nice study to some extent, less about, gut health, but not really.

This is a, a, a, a lovely study again from the, the horse microbiome Group. So this is Joy, la, with Chris Proudman, as the, as the PI on this particular project. And this is a really nice example of following a group of thoroughbreds that from foals into racing.

And looking at the bacterial community profile and looking at their athletic performance using various racing metrics. And one of the things that really stands out is horses that received antimicrobials very early on in life, often for respiratory conditions. These animals typically had reduced diversity in their gut microbiome, and this actually reduced their Their performance later on in life.

So if they had antimicrobials in that first month, you could predict later on in life they were less successful racehorses, and this just really shows how, how important and how fragile the gut is at that young state of life and actually they have that reduced diversity going forwards. So, you know, something as simple as some bacteria in the gut can have quite a large knock-on effect. I've talked about a lot of studies and a lot of stuff that I'm quite passionate about, but I suppose it always comes down to when we're thinking about looking at studies and thinking about practical application, what do these mean?

How is this really helping us manage horses today to improve, risks associated with gastrointestinal disease? And I've come up with this sort of list of, of, of take-home points. The question being, can the gut microbiome and metaboloma inform us about disease?

Yes, it certainly can, and we've seen lots of studies that show that we get differences in bacterial community profile and meta metabolite output that are associated with disease. But the next question is, well, is this ready to be used as a clinical tool? And I think that we still don't know enough about what's normal.

We don't know what a normal horse gut microbiome is because there's so much animal variability. It's a great research tool. It's really helping us understand the black box.

It's definitely helping us improve horse management and therefore horse welfare, but probably as a tool of its own at the moment, especially when we're just using the microbiome. Biome, which is very broad, I'm not convinced we're completely there yet. Hopefully one day we'll definitely get some biomarkers and some, you know, use of this as a diagnostic tool, but until we get to that point, I think it's a fantastic research tool.

It's definitely shaping the way that we manage our horses, but we're not quite there yet to use this as a sort of silver bullet to help fix colic. And that is an overview of using microbiome and metabolismome data to sort of help us further understand the last 15 years of research around intestinal disease in horses.