Glucose monitor misinformation, and AI dairy farms
In this edition of The Naked Scientists: Continuous Glucose Monitors are increasingly popular. But are they feeding us dietary misinformation? Also ahead: the “sexome”: scientists describe the genital microbiome, and how it might help with forensic investigations. And, we go to Wales to find out how artificial intelligence is helping dairy farmers...
In this episode
01:01 - Glucose monitors could feed you misinformation
Glucose monitors could feed you misinformation
Javier Gonzalez, University of Bath
Continuous glucose monitors are becoming all the rage: endorsed by celebrities, these devices resemble a coin-sized patch that sticks on the skin and beams a read-out of predicted blood sugar levels to an app on a mobile phone. Many people have signed up to programmes that claim they’ll help improve your health by guiding your nutritional intake. And although these devices can be very helpful for people with diabetes, where they can be reasonably accurately calibrated to the user’s blood sugar, the evidence is, according to a new study, when the average person straps one on, they may be being fed misinformation about the impact of their diet, potentially leading them to make worse food choices. Javier Gonzalez is at the University of Bath…
Javier - We were interested in the accuracy of these continuous glucose monitors, sometimes called CGMs, and these CGMs are devices that attempt to estimate our blood sugar level at any one time, and how they do this is they don't measure the sugar in the blood, they measure the sugar in something called interstitial fluid, and that's the fluid that bathes our cells and it's slightly separate to the bloodstream, and it then uses algorithms to try to estimate what the true level is in the bloodstream, and because of this difference we don't expect exactly the same value from the CGM to the blood sugar level, and because some influencers on social media and TV shows and also in research are tending to use these CGMs to try to understand the blood sugar responses to foods, so for example does a banana increase your blood sugar level more than another food, then in this scenario there are potentially changes in factors that influence the accuracy of CGM, and we wanted to better understand that.
Chris - I must admit I work in a hospital, I've had colleagues roll up their sleeves and show one of these patches, they look like a sort of sticking plaster stuck on that they've worn for a couple of weeks, and then they whip out their mobile phone and show you a trace saying - “Well, that was my breakfast and look what it did to my glucose level.” So, this is not just the odd social media influencer, this appears to be quite penetrating through society in terms of the uptake of these devices.
Javier - Yeah definitely, they were originally developed for people living with diabetes and can be a great tool for people with diabetes to manage their blood sugar level, but as you say they're increasingly used by the general population.
Chris - What did you do then to try to probe this? How accurate these things are compared with, and I presume that pricking your finger like a diabetic would, that's the gold standard because that's really blood?
Javier - That's right, yeah, so what we did is we used the finger prick test, the finger stick test, as our gold standard measure, and at the same time people wore a continuous glucose monitor, and what we did is we studied people across a variety of different foods that should lead to different levels of blood sugar, so we gave people pure glucose and then we gave people different foods and different fruit smoothies, sometimes it was whole fruit, sometimes it was in the liquid form, to try and understand whether the accuracy of the CGM differed across these variety of different foods and beverages.
Chris - How many people did you look at?
Javier - This was a study of generally healthy people, it was a sample size of 15, which for some types of studies can be relatively small, but what we did was a crossover design, so people act as their own control, and for those types of studies you can get much more statistical power for a very small sample size, so it was in some senses proof of concept, it's not necessarily aiming to extrapolate across a whole load of people, but it does demonstrate the principle in a tightly controlled laboratory setting.
Chris - And were they any good? Did the devices turn out to be an accurate reflection of what was really going on, or are we being misled?
Javier - The general finding was that on average they tend to overestimate the true blood sugar level, but what complicates this is that that difference was varying between people and between the different foods and beverages that were ingested, and that's trickier because if it was just a clear systematic bias then we could easily just subtract that and correct for it across everyone and across all the meals and it would be rectified, but because it varies between people and the different foods and beverages, then it's not easy to apply such a simple correction factor to improve the CGM accuracy.
Chris - Now, if these things were condemning bad for you stuff, you wouldn't mind if it told people that it was spiking their glucose higher, because it might deter them from resorting to that extra can of fizzy drink or that bar of chocolate, but was it condemning things that actually we quite like people to be eating more of?
Javier - Well it was certainly misclassifying certain things that we would describe as low glycemic index. So, the glycemic index is a way to classify foods by the degree to which they raise your blood sugar level, and it was classifying some foods as medium to high glycemic index when they're actually classified as low glycemic index based on the gold standard measure, and I guess one other thing that might be causing a bit of unnecessary concern is that because they overestimate the blood sugar level systematically, then it also means that when we're aiming to keep our blood sugar level within a certain range, the CGM overestimates the time we spend outside of this healthy range. So, you could have someone who has perfectly healthy blood sugar levels, but they're unnecessarily worried because their CGM is telling them that they have a higher blood sugar level than they actually have.
Chris - I was just going to say, is this going to feed into people's health anxieties, and is it actually going to make some people potentially eat less well because they'll be thinking that some things are worse for them than they really are?
Javier - Potentially, and I think that's something we should be aware of anyway, in the sense that we should be not focusing on one single metric. Our health is dependent on a variety of different things, and glucose is just one of those factors, so even if this was a perfect measurement, there are other things that contribute, and we might make dietary decisions that improve our blood sugar level at a detriment to other aspects of our health.
07:24 - Partners transfer microbial 'sexome' during intercourse
Partners transfer microbial 'sexome' during intercourse
Ruby Dixon & Brendan Chapman, Murdoch University
Researchers in Australia have found that sexual partners transfer components of their distinctive genital microbiomes to each other during intercourse. This sexome - as it has been dubbed - is unique, and it’s thought that it might prove pivotal in sexual assault investigations in future. Ruby Dixon and Brendan Chapman are researchers at Murdoch University, in Perth…
Brendan - We have had challenges for a long time in forensic science with our ability to identify perpetrators of sexual assault. What we end up with are usually samples from a sexual assault victim, which in most cases is a female victim, and we are looking for cellular material or biological material that's been deposited by the male offender in most cases. We're largely looking for sperm cells. They can be quite difficult to isolate. We find that we have the most opportunity as soon as possible following the event.
Chris - So, what do you think it might be possible to do about this then?
Brendan - So, given that window, any opportunity for finding biological material - or a trace beyond that window - is incredibly important. So, hence the research that we are looking at. The idea came more so from the ability to identify something that wasn't human material. And that's where we came to approach it from the perspective of a transfer of bacterial material or the healthy bacteria that lives on us.
Chris - Ruby, tell me more. How does this work?
Ruby - So, in this project, we kind of worked off the principle, which is quite well known in forensics, that every contact leaves a trace. And in this circumstance, we looked at the bacteria that naturally exists on the genital region of both males and females. And we wanted to see whether there was any bacteria that was unique to a person, and if that potentially unique bacteria could transfer during sexual contact.
Chris - Well, they do say that your microbiome is more unique to you than your fingerprint even. I mean, I think that's a slight exaggeration there. But that's sort of what you're saying then, isn't it? That the microbial makeup on one party will be conveyed to the other party during any kind of sexual contact, and it might be possible to pick vestiges of that up.
Ruby - Yeah, that's exactly right.
Chris - So how did you explore it? What did you actually do to test whether that might be the case?
Ruby - What we did was we recruited volunteers who were in monogamous, long-term relationships to collect penile and vaginal swabs both before and after intercourse. We then took those swabs, extracted the bacterial DNA from them, and run a bunch of sequencing to show us what kind of bacteria was present in each sample. And then we wanted to compare them between the individual themselves. So, whether we looked at the same individual both before and after intercourse to see how their bacterial makeup had changed.
And then we also compared them between couples and then across all of the volunteers in our cohort.
Chris - What did you find, Brendan? Did this lead to a shift in the genetic fingerprint of the microbiome after there had been sexual contact, indicating that there had been things conveyed from one to the other that you could detect?
Brendan - One of the first things that we really noticed was that there was a disruption to the microbiome or a change. So, then what we had to do is really kind of hone in on that and look at what that change was. And that's where, I suppose to use the term fingerprint, we were really looking for those unique signatures, those genetic traces that were unique to one individual that we saw before intercourse in that individual, and then to see the same trace in their partner after intercourse. And that was really, I suppose, the eureka moment where we were able to observe that there was this very, very specific signature that was transmitted through sex that we could detect.
Chris - Was it on both parties, Ruby? Did you get the same sort of exchange both ways? And for how long could you pick that up?
Ruby - So, to answer your first question, it did go both ways. But we found that there was more transfer that occurred from the female to the male. We also weren't able to determine how long after intercourse that those bacterial signatures stayed present, as we had only taken a single swab.
Chris - Do you think, Brendan, this is relevant to a one-off event, though? Because as you've said, you're talking to monogamous couples. So, won’t there have already potentially been some mixing of the microbiomes, and therefore they might be in a slightly different situation because they'll have selected out or changed the microbiomes between the two parties already because of the influence of each other? And were these two completely unknown formally to each other? And I know that's not exclusively the case and that people are raped and assaulted by people they know. But in cases where you don't know the person, would that not be something interesting to look at to see if the same thing occurs or you get an even stronger signal under those circumstances?
Brendan - You're absolutely correct. And my response has really gone back to the point that we've had to start somewhere - this is the starting point. We suspect that because of the long-term nature that these couples have been together, then there may be some sort of sharing of some bacterial types. But really at the moment, we've kind of scratched the surface of understanding how this works and what happens. And as you would know, Chris, with any research, a lot of the time the results just open up more questions. And those questions are great because they drive the next phase of this research.
Chris - What's the next step, Ruby? Where are you going next?
Ruby - We want to go back and potentially target those bacterial species that we know are common in the female vaginal microbiome. And we want to be able to figure out a methodology that we can use to target those specifically rather than having to sort through an entire puzzle, I guess, of different bacterial species that might not be relevant to sexual assault casework.
Why we scratch an itch
Daniel Kaplan, University of Pittsburgh
Scientists in the United States say they have unpicked the real purpose of scratching an itch: scratching activates pain nerves, which turn off itch nerves but turn on immune defences in the region of the irritation. The resulting feedback loop is thought to have evolved to protect us against infections. Here’s Daniel Kaplan at the University of Pittsburgh…
Daniel - Cells of the immune system can release small molecules that are responsible for triggering some of our sensations, like painful sensation or itch sensation. A number of years ago, we looked at a type of neuron that encodes pain sensation and found that if we just activate these neurons by themselves, that's more than enough to trigger all kinds of inflammation in the skin. So we were curious to continue that kind of work and look whether something similar happened with neurons that encode itch sensation.
Chris - That's intriguing. So when you activate the pain nerve, it also then turns on immune responses, is what you're saying, and you're wondering, well, what about itch then? Does an itch turn on an immune response as well?
Daniel - That's exactly right. What we found though is that activating the neuron itself didn't seem to trigger any inflammation. Rather, what happens is that when the neurons that encode itch sensation are activated, you get the expected result, which is the mice start to scratch. But the unexpected finding was that it was the act of scratching that was actually required for inflammation in a number of different models of eczema.
Chris - That's very interesting. So there are specific nerves that just tell your body, this patch of skin itches, and there are another class of nerves that say this patch of skin hurts. But when you activate the itch ones, you feel an itch, but you don't get an immune response. You activate the pain ones and you do get an immune response.
Daniel - That seems to be true, at least as far as we've been able to determine.
Chris - And so your hypothesis is, well, when I itch, I scratch, the scratch causes some pain and the pain causes some inflammation.
Daniel - Well, that's actually what we discovered, is that the act of scratching ends up activating these pain sensing neurons, and that then leads to inflammation in the skin. So in that sense, by scratching, it actually sort of completes this circuit such that now you get nerve-driven inflammation to something that otherwise wouldn't cause pain.
Chris - What cells are actually releasing the immune effectors then, these responses, these signals that you're saying you get when you evoke pain and it produces an immune response? Where's that coming from? What's the circuitry there?
Daniel - We know there are a number of these small molecules that are released from the neurons, and some of them can trigger a type of cell in the skin called a mast cell. These cells have a lot of pre-made chemicals in them, and when they're activated, these chemicals get released into the skin, and that leads to inflammation that's important for the response to bacteria.
Chris - Mast cells are jam-packed with histamine, aren't they? Because when we have an allergic reaction, you get a histamine response, and that causes swelling and itching. So doesn't this feed back on itself then? If you've got a cell that releases something that makes you itchy and it was the itch that made you scratch in the first place, doesn't this mean you end up scratching yourself to death?
Daniel - That's absolutely correct. So when you have scratching, the mast cells are activated, this does release histamine, the histamine then acts on receptors which are on these itch-sensing neurons, and you get then more itch. This is one of the reasons why we have this phenomenon called the itch-scratch cycle. So if you scratch an itchy area or a rash, it feels very good, but then a few minutes later, it itches significantly more, and that's part of this positive feedback loop that gets developed.
Chris - In one respect then, that's a good thing, because if you're detonating immune responses and inflammatory chemicals from your mast cells like this, that's going to put the immune system on high alert. So if, say, a mosquito or something, we're trying to drill holes in your skin, you've been alerted to the fact that that patch of skin is not in a good way to pay attention to it. But that also sounds like it could have negative impact, because if you have this itch-scratch cycle, what breaks the cycle? Because otherwise, you would literally dig holes in your skin, wouldn't you?
Daniel - Well, this is part of the paradox of scratching. So on the one hand, scratching alerts you to something that needs to be removed from the skin, and from our findings that the scratching by activating these pain-sensing neurons, triggering inflammation, has a very important role in limiting bacterial infections at that site. But at the same time, it makes the rash a lot worse. So in general, we scratch an itch, and it often feels pleasurable, this common sensation. However, we will scratch an itch until it starts to hurt, and that's usually the trigger to stop scratching. So that's one of the ways in which the itch-scratching cycle can be broken in a short time scale. But for many people, this itch-scratch cycle becomes quite a problem and requires medical intervention.
Chris - So that would be someone with eczema or something like that, acute dermatitis?
Daniel - Yeah, precisely. And there are many forms of dermatitis, and the scratching is well known to make the eczema rash significantly worse, oftentimes to the point of needing a visit to your dermatologist to help you sort it all out.
Chris - Presumably then, a better understanding of this sort of wiring that's going on, as well as the interaction with the immune system that you've uncovered with this, that does give us some possible new insights into ways that we might be able to better manage things like dermatitis.
Daniel - Yeah, that's absolutely correct. In fact, previous research from a number of other people have found that the molecules released by the types of immune cells that are associated with allergy can themselves cause itch. And in fact, there are a number of new medications that target those molecules, which have been incredibly helpful.
20:49 - AI provides personalised cooling for dairy cows
AI provides personalised cooling for dairy cows
Andrew Gardner & Chloe Rodriguez, Galebreaker
This work was carried out with the support of UK Research and Innovation.
Cow’s milk is an integral part of most of the UK’s daily diet. But - much like other parts of the food sector - it is coming under increasing pressure from climate change, and these challenges mean we will need to inject a bit of cutting edge technology into the tried and tested traditions of farming. Will Tingle has been to southeast Wales to see this for himself…
Will - There are around about 1.9 million dairy cows in the UK right now. Given the cost, space requirement involved, and emissions produced by a cow, it does make sense to get the most milk you can from each individual, rather than increasing the size of the herd. Now much like us, a stress-free cow yields a better and more bountiful product. And one thing that cows really do not like is excess heat stress, and that is something I can entirely relate to. So how do we adapt to the behemoth of climate change whilst increasing food production for our growing population, and doing it in a sustainable and ethical way? That question has brought me here to Usk, a town in the east of Wales, to witness the exploits of agricultural ventilation experts, Galebreaker. My guides today are their animal welfare specialist Chloe Rodriguez, and kicking us off, Technical Director Andrew Gardner.
Andrew - So Galebreaker has always been about natural ventilation in farm buildings, so bringing fresh air and trying to create the optimum environment for the animal. One of those with the dairy cow is actually heat stress. So yeah, this project was all about how we can embrace artificial intelligence to help with that, solving that problem.
Will - Chloe, to bring you into this, heat waves and heat is becoming an increasing problem around the world. What does heat do to a cow that is a problem?
Chloe - When you're looking at dairy cows and heat stress, you're looking at a combination of the temperature and humidity, and it's actually because on a day that we might think, oh it's a lovely day, put our t-shirts on and it's nice and go outside about 18 degrees. Actually, if the humidity in the building is up at 70 or 80, cows could be getting quite heat stressed at that. That humidity, that lets cows be heat stressed at lower ambient temperatures. Cows cannot regulate their temperature like we can through sweating, so the only way that they can really regulate their temperature is through increased respiratory rate or panting. And again, when a cow lies down, she accumulates heat and when she stands up, she can get rid of her heat, so that's why on a hot day you'll see a lot of cows standing.
Will - If the heat increases by a considerable amount, you're expecting to see the cows suffer physiologically?
Chloe - Yes, so initially you'll see behavioural changes, the cows will have increased standing times, they'll be panting, and then in turn that will affect fertility, that will affect lameness rates, and then eventually that will also decrease milk production.
Will - Do we have any, and I hate to be so callous as to put a number on it, but do we have any idea just how much heat can affect the cow?
Andrew - During the first phase of this project, where we were actually monitoring animal behaviour during the summer of 2023, there was one significant heat stress event where not only is the cow dropping its milk yield during the heat, but because of the stress that they've incurred, we actually found that their milk yield dropped for the next 20 days following that heat stress event, and actually yeah, during that time, we saw on average those cows dropping 50 litres of milk, what they could have been producing if they hadn't seen that heat stress event, and the price of milk that they were getting for it, the losses on the farm was about £2,500.
Will - And with losses like that, it really is in everyone's interest to keep the cows cool, and this is where the forefront of AI is allowing us to consult perhaps the leading authority on how hot a cow is feeling, that is to say the cow itself. It's time for myself, Chloe and Andrew to take a trip down to the cow shed and see what the future of dairy farming could well look like.
Well we are now deep in the cow shed, this smell brings me back to my youth. What I'm looking at here is I'm looking at a couple of vent tubes with holes in the bottom aimed down towards some resting cows, and I'm looking at some very fetching translucent shades on the outer walls of the building as well. Talk me through what these are.
Andrew - Yeah, okay, so on the side of the building we've got the side curtains, and it does allow natural light through, but as it gets warm we want that curtain to open so it rolls up, allowing fresh air into the building. Above the cubicle itself though, where we want the cows to be lying down in those hot conditions, just like a human, what we like if we're in an office we want to have a desk fan or a stream of air coming across us to help us keep us cool. So what we've developed is a positive pressure tube, so that's that inflatable tube that goes directly along the length of the cubicles, and there's a jet of air that's spraying onto the back of the cow keeping it cool, and that's where their heat production is around their rumens. For positive pressure tubes in the past we were literally advising the farmer to switch them on when they think it gets hot and leave it running continuously. With regards to the curtain we were measuring the temperature inside the building and opening and closing it. We want to put the animal first in that control system and therefore if we can monitor the behaviour of the cow we can let the cow decide, or the cow's behaviour decide, what does the cow need and therefore which of our ventilation systems need to be activated.
Will - You're essentially saying then we're allowing the cows to maybe not choose but dictate how and in what way their building is ventilated to keep it cool. Why is this different or why is this better or more beneficial than the old way?
Andrew - Every cow is slightly different and the conditions they're experiencing are quite different. So in the old way of where we've just got a single temperature sensor we basically treat every cow the same, therefore some cows might have needed that ventilation or cooling earlier and the thought was that we would create a more optimum environment for each and every animal.
Will - This is far more customisable, is it also perhaps slightly more energy efficient because you don't have to have constant mechanical opening closing that sort of thing?
Andrew - No exactly, so the fact that when you're treating the herd as an average or on everything then actually it's probably more common that you would be running the controls or running the devices for longer and the cost of running a fan can be quite sizable certainly with the increase in electric costs.
Will - Big question then, how are you getting all of this individualised personalised cow data?
Andrew - SmartBell had an ear tag that goes onto the cow which has an inbuilt accelerometer and an ear temperature sensor. The tag is monitoring that ear movement and really subtly that ear movement can be different if they're eating, if they're lying, if they're ruminating, if they're drinking, walking and therefore it was through a machine learning approach where we could use that ear movement to tell us what each and every cow was doing. So we were effectively monitoring each cow to know how long they're lying down, how long they're eating, is that normal, is it dropping off and we saw how those behaviours correlated with heat stress events and therefore we can activate the control system earlier.
Will - So this gives you much more concurrent data as opposed to one single temperature reading in the centre of a barn?
Andrew - Yeah exactly.
Will - Can I have a look at one?
Oh here we go, this is tiny, there must be so much information being packed into this and yet we are looking at something smaller than a padlock on the cow's ear and presumably this has no real effect on the cow's day-to-day activities?
Chloe - No not at all, cows have to be tagged anyway so they wear two large earrings anyway so it's just another tag in the ear.
Will - Where does this information go?
Andrew - The tag is Bluetooth connected to a receiver which then sends up to the cloud so everything is up in the cloud, all that data is being collated remotely and we can access that data at any time as well to see exactly what cows are doing, individual cows are doing and their general behaviours. That's how SmartBell was obtaining all that data to then do the subsequent analysis.
Will - So this is all very impressive, where are we going forwards from here?
Andrew - We certainly wouldn't for the cows sake hope for a hot summer but our intention is for 2025 that we can actually really validate if this system works and then with our true validation really seeing the improvements in animal welfare, production and a happy animal and a happy farmer then we would really feel we've got a commercial proposition to take this wider.
29:48 - Why does lactose intolerance vary across countries?
Why does lactose intolerance vary across countries?
Garth writes in asking about rates of lactose intolerance around the world. James Tytko asked Mark Thomas, Professor of Evolutionary Genetics at UCL, to help with the answer...
James - Thanks Garth. Lactose is a sugar found in milk. The enzyme responsible for breaking down lactose into more absorbable sugars is called lactase, and it’s something we make a lot of as babies when, for the most part, our diets are significantly constituted of our mothers’ milk.
But some time after the weaning period, most of us stop producing this enzyme, a genetically determined trait known as lactase non-persistence. This is what causes people to experience lactose intolerance.
Here to elucidate a bit more of the detail - and, here’s a warning, it’s a bit grisly if you’re having your lunch - is Mark Thomas, Professor of Evolutionary Genetics at University College London…
Mark - So, if we’re not producing lactase, then lactose will move into our large intestine, causing fluid to go out of the blood, into the large intestine, liquifying poo, causing diarrhoea. The lactose can also be fermented by gut bacteria causing bloating and flatulence. Those are the symptoms we associate with lactose intolerance…
James - Now, to your question, Garth, is this trait as variable among populations in different countries as you suggest, and if so, why?
Mark - Continued production of lactase into adult life is very common in Europe, in very high frequencies in Britain, Ireland and Scandinavia. It's relatively rarer in populations in East Asia, native Americans, native Australians, and others. Why has it got that lumpy distribution across the world? The reason is that the genetically determined trait of lactase persistence has co-evolved with the practice of dairying. We have been dairying - domestically keeping animals and extracting milk from them and consuming it - for about 9,000 years. Because milk entered the diet of adults, the trait of lactase persistence, being able to extract the energy from that milk, has co-evolved. This is what we called a gene culture co-evolutionary scenario. In fact, it's probably the best example we have in humans. Lactase persistence developed in places where we were dairying early, like the Middle East, Europe, many populations in Africa and also in Southern Asia. In other regions around the world, although there were other animals that were domesticated and could be milked (like lamas and alpacas) they generally weren't.
It's actually co-evolved under very, very strong natural selection pressures. In fact, for a single gene trait it's the strongest signatures of natural selection for the whole genome that we see in many populations around the world.
Comments
Add a comment