Quirks and Quarks - Cocaine in waterways makes salmon roam further, and more…

Episode Date: May 1, 2026

Cocaine and many other chemicals and drugs are found in many waterways, but especially around wastewater treatment plants. Scientists exposed wild juvenile Atlantic salmon to cocaine and its byproduct to see how it impacted their behaviour in the wild. As a result, the fish swam twice as far, which could put them in more danger.

If you sold somebody a loaded gun who you knew was in a vulnerable state and they shot themselves. I think it is murder. Just because you're using the internet doesn't mean you get away with murder. I'm Damon Fairless, host of Hunting Warhead. This season, I take you inside the business of suicide, and the places desperate people go when they can't find what they need in the real world. Hunting the Suicide Salesman. Available now, wherever you get your podcasts.

This is a CBC podcast. Hi, I'm Bob McDonald. Welcome to Quirks and Quarks. On this week's show, what an AI scientist means for the future of human scientists. I think the real story here is just how soon and how quickly AI will be able to do science on par with the top human scientists or even better. And cocaine pollution makes salmon roam farther. They swam one and a half times further.

and then by week seven they were swimming almost two times farther. Plus, new insights on the origin of life, horses can smell when we're afraid, and looking back at scientists blasting music to plants. All this today on Quarks and Quarks. Artificial intelligence is undergoing a seismic shift in science that's shaking the very foundations of the scientific process. For the first time in history, scientists have all of humanity's knowledge at their fingertips, and the power to wield it in previously unimaginable ways.

Like how Rutgers University mathematician Dr. Lisa Carbone found out when she put a paper she'd been working on for several years through Google Gemini 3 Deep Think to fact-check it. It came straight back with, no, that's not correct. It gave three separate irrefutable reasons why our mathematical arguments around one particular, the statement were incompatible. This was pretty destabilizing because the paper had already been peer-reviewed.

I debated, and the model didn't try to appease me, as most AI models do by trying to guess what you want to hear. It took me a while to understand because it was really outside of my thought process, and the model's reasoning was completely correct. Naturally, AI developers, like OpenAI CEO Sam Altman, have a rosy-eyed view of what lies ahead. We now, for the first time in the last few months, have made AI systems that are smart enough that our best scientists are saying, this changes the game for science. This means science will move faster.

These models are discovering new knowledge on their own. AI for science has expanded beyond the simple pattern recognition and prediction capabilities of earlier AI systems to take on part or all of the scientific process itself. from coming up with new hypotheses to conducting a kind of peer review. Here's Dr. James Soe, who leads the AI for Science Lab at Stanford University. I think we're in the midst of this really exciting paradigm shift, from thinking of AI as a tool for science, towards thinking of AI as a scientist by itself.

That's right. Artificial intelligence may be coming for scientific jobs, too. Here's what Harvard University longevity expert Dr. David Sinclair said about it in an interview with Peter Diamandis on his podcast, Moonshots. I thought that my job was not at stake. The arrogance was, I've got all this knowledge and experience and got feeling, and I'm really creative. But here I'm seeing the beginnings of creativity that can be super creativity in the future, and it'll only get better. But as much hope as there is about what AI can do for science, there is concern. Concern about AI not being good enough yet and making stuff up. Concern about fake data getting into the scientific literature.

And concern that immensely powerful AI scientist technology could be wielded against society. So, will AI and science save or destroy us? In a recent paper in Nature, Dr. Jeff Klune and his colleagues recently developed an AI scientist pipeline to automate the entire scientific process, from generating an idea to doing research and even publishing and reviewing its own paper. Dr. Clune is a professor of computer science at the University of British Columbia and a C-FAR AI chair at the Vector Institute who used to work at OpenAI. Hello and welcome back to our program. Thank you very much for having me. How close would you say AI scientist systems are compared to what actual human scientists can do?

So the paper that we just had in nature a few weeks ago show that AI can do the entire arc of producing a scientific paper, from coming up with its own original idea to figuring out an experimental plan to test that idea, to creating and coding up those experiments, running them, looking at the results, visualizing the results in charts and plots and then writing an entire paper summarizing the entire process and even peer reviewing it and saying, you know, is this any good? And if it is good, coming back to the human and saying, I've done a scientific experiment and I've written it up in a human readable format. And we submitted with that paper to a peer review process at a top tier machine learning conference and ended up, it scored better than 55% of human authored papers and was accepted via the peer review process.

Holy smokes. Yeah, it's pretty incredible. So this is evidence that it can do it. Now, your question is how well can it do it? My answer right now is that it's about as good as a first or a second year graduate student. But I actually think it's a distraction to focus on the current capabilities of the system instead of recognizing the trend line, which is how quickly AI in general and AI scientists in particular are getting better.

The writing is on the wall. These systems are getting better. amazingly quickly. And in AI and machine learning, one thing that has become clear is that once something barely starts to work a few years later, it's jaw dropping the good, if not superhuman. And so I think the real story here is just how soon and how quickly AI will be able to do science on par with the top human scientists or even better. So it's just a graduate student now, but soon it'll be a PhD. It's growing and learning really quickly. Yeah, maybe beyond that, it soon becomes one of the best professors in the world, or maybe it becomes super

intelligent, better than Einstein and Newton. We don't know where this is headed. So just fundamentally, I'm not an AI expert, but how is it actually doing that? How are these AI systems behaving like scientists? Great question. Yeah. So, you know, anyone who here who's interacted with chat GPT or any one of the frontier models knows that you can kind of ask them questions in English and get back answers in English. And so you could tell a system, you know, please look at this area of science and propose a new research topic that should be studied in this area. And it could use its knowledge from having read the internet. Maybe it could do web searches and read about that area of science.

And it would come back and it would propose a few ideas to you. And then you could say, all right, pick the idea that you think is best and come up with an experimental plan. And it would think about it for a while and it would come back and give you a set of experiments. It thinks it should run. And then you say, okay, you know how to code. Why don't you code those experiments up? and run them. And if those experiments are like AI or computer science experiments, it can run those experiments on the computer. And then you can say, why don't you look at the data and see if the data

confirm the hypothesis or you want to run new experiments, and it would do that. And then eventually, like, why don't you visualize the results in charts and plots, whatever you think is best? And now, could you write an entire scientific paper that summarizes these results? That gives you a high level flavor of how you can use the technology that you're familiar with and a series of requests to have it kind of go through all of the steps that a scientist would do. It's a little bit more complicated under the hood, but that's basically what we're doing. Boy, how long does all of that take? Once the system is really good, you actually can determine for each of those steps, how long should it think about it. But roughly speaking, it can generate a whole paper in minutes to

an hour or two hours. But that's astounding. It can do that in minutes or hours. It takes scientists. is sometimes years to do stuff like that. Here it is doing it so quickly. That's the general lesson of computers, right? If you go back to the time before Turing, you had entire rooms of people who were called computers doing math operations and calculations and running algorithms by hand, and now your smartwatch can do that. Now apply that same lesson to all of human scientific discovery, and you can see where this is headed. So what are some of the major advances in artificial intelligence since the development of large language models like chat GPT that are paving the way to this automatic scientific process?

Yeah, so one of the most important things that has come onto this scene is what are called reasoning models. It used to be the case that no matter whether you asked what, you know, time it was or what the weather is today or, you know, for it to come up with a novel mathematical proof that solves one of the great open challenges in mathematics, it would run the same amount of computation and it would take the same amount of time. it would just give you the best answer. It didn't know how to stop and think really, really hard about a really hard question

and spend a lot more time thinking about it than a very simple question. But a few years back, OpenAI came up with the innovation of what are called these reasoning models. And now, if you ask it, I would like you to go off and think of a really, really important new scientific direction to pursue or try to solve one of the great open challenges in mathematics or physics. It will spend a lot more time doing that. And it internally is actually acting in many ways like a human would do. It comes up with kind of maybe like a big plan of action. Like first I should do this, then I should do this, then I should check my thoughts against like the internet.

It has these complicated workflows and internal steps it can do. And so it can think really hard and come back and give you an answer. Another big development is what's called multimodality. It used to be the case that these systems were only text in, text out. But now increasingly they can look at images, they can look at video, they can do web searches. and read the results of web searches. Sometimes they can even browse the internet and look at the actual pixels on the screen.

So they're getting kind of more of the senses so they can see and understand the world better. And that allows, for example, in our paper, for them to do an experiment and see the data and then chart the data, much like a human would do, and get a really good quick sense of like how that, you know, experiment is going rather than like staring at all the individual numbers and trying to make a sense of the forest for the trees.

You're talking about the computer thinking about things. I mean, how well can these AI models actually reason rather than just come up with pattern recognition? Yeah, I actually think it is a disservice to humanity to try to dismiss what these models are doing as simple pattern recognition, which is what a lot of people like to do. In my view, there's a long arc in human history of us wanting to feel really, really important. We did not want to accept the fact that we weren't the center of the universe. We did not want to accept the fact that we were related to other animals and have a common ancestor with things like chimpanzees and bonobos. We didn't want to accept the fact that we weren't the only intelligent thing that could do things like

play go and chess, these sorts of games. And currently, we don't want to accept the fact that we're not the only thing that can do thinking and reasoning. So often when we see an AI, even if it's coming up with new scientific ideas, new scientific advances, we like to come up with excuses to make us feel better and say, oh, that's not real thinking, that's not real creativity. that's something else like pattern recognition. But I actually think, while the systems are not as good at a human in all domains of creativity and thinking,

what they're doing genuinely is thinking and its reasoning. It's just not as good as us yet, but it will be better than us soon. And so we should stop and have a real recognition that we're finally meeting the first other intelligent entity in the universe that we've ever encountered. That we created. And we created.

In one study, Dr. David Sinclair, the longevity expert from Harvard, developed a system that they fed their data into. Here's what he said. What was incredible was it didn't just come up with validating what the field, the smartest people in my field, had done over the last 10 years. What happened was it came back and said, hey, did you guys ever think of this before? And came up with a completely new way of looking at the data and making a new model to predict biological age out of the data we gave it. Not only that, it proved the data, it did the statistics, it wrote the paper up for us and presented us with the finished product, which sucked because we want to be co-authors too. So Dr. Cologne, what kind of limitations do these AI scientist systems currently have? You know, so their ideas sometimes are creative but aren't always that creative.

their implementation of the ideas and their kind of scientific process, the rigor with which they go through the scientific process, is okay but not great, which is also true of entry-level graduate students and undergrads. Their ability to look at the experimental results and judge whether they really confirm the hypothesis, you know, sometimes the models get overconfident and think they've kind of proven their case when they haven't. That happens sometimes. And sometimes they'll do boneheaded things that we know AI systems do, which is they might just like completely make up a reference that doesn't actually exist. And so part of the work that we did was trying to like increase those types of skills in the system by creating feedback loops where we like ask it to make

sure for every reference that it puts in the paper, that it triple checks that it's a real reference and points us to the actual paper. So it reduces the kind of mistakes that it makes. So some of the scientists that you just played quotes from are saying it is helping me feel. figure out bugs in my reasoning. It will be able to figure out bugs in its own reasoning. And it's making new innovations that we hadn't thought of before. And then it's doing the experiments to show that those ideas actually work. And that is just the beginning of what I consider to be like the dawn of a new era in science in which AI-driven scientific progress is extremely powerful, extremely rapid and extremely dramatic. And there are risks to that,

But there are also tremendous upsides. I think it is possible that we will solve all disease, primarily driven by AI scientists rapidly making scientific progress and helping us eliminate and understand the biology behind disease and then come up with the cure. But at the same time, we're already seeing the emergence of AI generated fake papers and fake data. I mean, how big of a problem is that in this generation of science that's run by AI? That's a really good point. I think that is going to be a real problem. I suspect it probably is currently a real problem that we haven't quite fully realized the extent of yet.

However, I also think it's probably a transitory problem. I think it'll be a few years where you get a lot of AI slop, but soon the AI itself will be really good at judging that kind of chicanery and like the garbage science and basically calling it out. There's actually an opportunity here. I mean, science is very expensive to do. So very often we have a human team. They come up with a result. Either the result is due to chance. It was luck or the human team itself maybe even fraudulently invented the result.

They publish it. And because it's so expensive to replicate that pit of science, nobody else goes and checks whether or not that it's actually like real. With AI scientists, because they're automated, we actually have the ability for every paper that gets published and claims that this method or technique or drug has this benefit, the system itself can actually go and rapidly try to replicate the science and find out if it's true. Now, that's really easy for computer science experiments or experiments you have automated laboratories, like a robot that can kind of mix and match chemicals and materials and do biological experiments or

chemistry experiments or material science experiments. Obviously, much harder in things like human trials. But basically, human scientists are incentivized to get promotion and tenure and fame and fortune. So they always want to do their own new thing. want to spend their life replicating the work of others to kind of check if it's real or not. But we could actually ask our AI scientists to do that all the time. And we might be able to reduce the amount of kind of noise in the system as opposed to having it be accelerated. Anytime a new tool comes along, there's always someone who finds a way to use it for nefarious purposes. So what about the potential safety issues with these AI scientific systems getting into the hands of the wrong people?

Yeah, this is one of the things that worries me the most. And it doesn't even have to be AI in the hands of a person with malintent. It could also be you could incentivize a scientific system to do things that end up making it do unsafe things. But one thing you could imagine is that like human scientists, we want to train these scientists to have a sense of ethics, to know that some experiments are not worth conducting. You shouldn't try to make a more dangerous version of coronavirus. you shouldn't do harmful things to children, et cetera. But we also don't want to trust the ethics of individual AI scientists, just as we don't trust the ethics of individual human scientists.

We also have oversight. So we need to build lots of oversight to make sure that the system, what is doing is safe, is ethical, and is beneficial for humanity. Because if we can make the system beneficial for humanity, there's no telling the tremendous benefits we can produce. But in the hands of the wrong people, or in the hands of the wrong kind of high-level objective, it might end up doing something dangerous.

So we want the system to shut down unethical experiments, whether they're requested by a human or proposed by the AI scientist itself. Well, with all the progress of AI scientists and how they're getting so good at doing everything themselves, what role do you see for scientists in the future as we continue down this path of letting the AI systems do everything automatically? Great question.

One thing I definitely think will happen is that the humans will be very important in terms of setting the goals and the objectives. Like what is the system working on and what kind of things is trying to learn about and explore? Much like a program officer at a funding agency or the professor of a very large lab would do or the CEO of a company. And so we can kind of help tell the system what we're interested in having it do. A second thing is what we just talked about, which is oversight of the safety and safety. the ethical issues at play. So we want to make sure that the system is doing ethical science and a science that benefits humanity and nothing that is dangerous or that would help a bad actor, a bad human actor accomplish some evil intent. And so I think there's going to be a huge amount

of human involvement in trying to kind of oversee the system and channel it towards positive outcomes rather than less positive outcomes. But the core job of most scientists now, like all economic areas of society may be automated by AI. We are staring at one of the biggest transformations in human history, maybe the biggest transformation in human history. It's right around the corner. And I don't want to pretend that it's not going to be disruptive. It is very possible that my job goes away as a result of AI becoming better at science than I. And my job transforms into being somebody who oversees these systems and make sure they're doing the right thing rather than doing the science itself.

Well, always keep humans in the loop. Yeah, I agree with that. The stakes could not be higher. We're talking about everything from solving all of disease, solving poverty, giving everybody wonderful education and a life of abundance to the other extreme of wiping out human civilization. And so there is nothing more important for humanity to be doing right now and trying to make sure that the development of superintelligence goes well.

Dr. Clune, thank you so much for your time. Thank you. It's my pleasure. Dr. Jeff Clune is a professor at computer science at the University of British Columbia and C-FAR, AI, Chair at the Vector Institute. And as we continue to celebrate 50 years on the air, we want to take you back to one of our favorite interviews from our very first season. Back in 1976, our first host, David Suzuki, interviewed a botanist from the University of Ottawa named Pearl Wineberger.

She was working on some unusual and even slightly controversial science at the time, playing loud sounds to seeds to see what that would do to their growth patterns. And to everyone's surprise, the seeds exposed to the music ended up growing twice as tall as the plants that weren't bombarded with these sounds. It's the type of science we've always loved to cover since day one, the research that explores the wonderful and quirky ways that the world works. Here's Professor Pearl Weinberger speaking with David Suzuki on our episode, which aired on February 4, 1976. What got you involved in this kind of effective sound on plant growth?

Around 1968, word came out of India that a Dr. Singh had been playing his violin to rice plants, and it was reputed that the rice plants increased the growth. And I was sitting around the table with the children, and together with them, I grinned. And then I thought to myself, here am I at a university. giving classes to first year students and saying to them that one of the things, one of the important things that they should get out of all of this training

is an objective way of looking at things and if they, not to dissociate themselves from any novel or avant-garde techniques, but go out and test it before they negate it. And so, having given myself this little admonition, just this. So how did you go about setting up your early experiments then? Could you describe what you did

and then the results that you got? I made use of friendships in the scientific community in Ottawa. I didn't have any oscillators at hand, and I phoned a couple of people who I knew and asked if they could loan me a couple, with speakers that would cover the correct range. Now, I was in my 40s, and I guess I'm rather deaf, because the upper limit of my hearing was 12,000 cycles per second, 12K.

So I thought to myself, well, one frequency that I will use will be my upper range, and just by haphazents, I tried 12K, since I tried 12K. So I used 5K and 12K. And this is one of the examples where Lady Luck played an enormous part in the experiments.

The next decision I had to make was how loud. And I thought, well, if this is going to work at all, I better blast the seeds that I was going to grow. And I used something between 90 and 100 decibels. That's equivalent to a jackhammer. Then, time of experiments. and what seeds. I'd been working with a winter wheat which requires a chilling period of six weeks to accelerate growth to flowering. I knew most about the growth characteristics of this wheat.

So I thought, well, I'll use this reed-o wheat and blast it with sound during its chilling period and see what happens. And I did just that. And to my amazement, the two groups of seeds grew more quickly than we'd ever shown control groups to grow. How significant were these differences if you had the non-sound-treated seeds and the sound-treated seeds and grew them up? What was a difference? You were measuring height, I assume. We were measuring, and I'm saying we,

because at this time, a young lady, Mrs. Measures, came in as a grad student and I put her on this problem. And we looked at changes in numbers of leaves, wet weight and dry weights of tops, that is the greenery and the roots. Under those conditions where the grains were exposed, both during the chilling period and subsequently,

that there were more than two and a half times increase in wet weight, and that number of extra tillers were formed too. Pearl, I would think that you've been subjected to a lot of ridicule by your peer group. Would you say this is true? Yes, there's been a lot of ridicule associated with us, but because of this, unfortunately, a breakthrough in our capacity to increase growth yield

has not been followed through. Yes, I'm always struck in looking historically at various scientists who stood at the forefront of their area that often when people made a discovery that was so far out of the accepted wisdom of the time, they're ridiculed as being fools or charlatans or what. And I think people don't realize, you know, the idea of a scientist is that we seek after truth.

But in fact, we're really very hidebound in our own prejudices and often will take steps that will prevent other people from doing really novel types of research. We all tend to think that the modern world, with all its advanced technology, has something so far superior over ancient wisdom, and some of which have become myths. I mean, obviously we can't accept all of this,

but certainly we shouldn't discard it all. That was botanist Pearl Wineberger in an interview with David Suzuki on our episode from February 4, 1976. I'm Bob McDonald and you're listening to Quirks and Quarks on CBC Radio 1 and streaming live on the CBC News app. Just go to the local tab and press play wherever you are. Coming up later in the program, a tiny molecule is helping us solve the mystery of the origins of life. Discovery in science obviously is always super exciting.

And I think what makes it interesting and rewarding is that discovery often comes from corners where you don't expect it. I am an actor, fresh out of theater school with big dreams and an even bigger drug habit. But things are pretty good. That is until my best friend is set up on a date with David Lee Roth. Yeah, from Van Halen. If you know, you know. From CBC's personally, this is Discount Dave and the Fix. The true-ish story about how a fake rock star led me to a real trial that held up a mirror to me.

And okay, let's just say that not everyone in this story is who you think they are. Personally, discount Dave and the Fix. Available now on CBC Listen or wherever you get your podcasts. You've heard of cocaine bear, but have you heard of cocaine fish? It might sound like a silly movie plot, but this is a very real environmental issue. Cocaine is just one of the many illicit and pharmaceutical drugs that end up in our waterways. This is a problem at home and abroad. According to a statistics Canada study, the city of Prince Albert has the highest amounts of cocaine, meth and amphetamies in its wastewater compared to other cities across the country.

It turns out there's a large range of well-known pharmaceuticals that are ending up in wastewater and making it into freshwater and marine environments. So these include common blood pressure medications, heart medications, even opioid, other pain killers, antibiotics, antifungals, antidepressants. Once the drug gets into the environment, oftentimes it can stay there and get into fish, according to a recent study out of Ontario. Fish living in rivers downstream of wastewater treatment plants are being exposed to more than just runoff. A new study from the University of Waterloo has found that antidepressants, opioids, and other drugs are accumulating in their bodies. It is the first time that scientists have documented how these drugs are distributed in wild fish in Canada. So what does this exposure to this cocktail of chemicals mean for fish? Well, it's a really hard question for scientists to answer.

But in another new study out of Sweden, a Canadian scientist has come up with a clever way to test how these drugs and their byproducts in hotspots around wastewater treatment plants might be impacting the behavior of wild fish we care about, like salmon. Dr. Aaron McCollum is that scientist. She's an associate professor of aquatic ecology at the Swedish University of Agricultural Sciences in Umeo, Sweden. Dr. McCallum, welcome to our program. Thank you.

So first of all, what were you hoping to find out in this study about cocaine and fish? So cocaine is an interesting compound to study because it's a common recreational drug. It's used very widely around the world. And we know that it has effects on human behavior. And since we knew that cocaine is being measured in the environment, we then wondered how cocaine and its main metabolite, benzalekagin, how they impacted Atlantic salmon swimming behavior in the wild. Metabolites? What's that? Tell me about it. So a metabolite is what is formed by the body. And it's basically a byproduct or a breakdown product of our bodies metabolizing so that it can be clear. cleared from our systems. So it's actually levels of the metabolite that are measured in higher

concentrations out in the waters than cocaine itself. So I'm surprised that the metabolite can still have an effect like the cocaine itself? Yes, exactly. So the metabolite is known to be a vasoconstrictor, which means that it constricts the blood vessels. And so it contributes a little bit to the physiological effects that we experience if you take cocaine. How did you go about testing how these drugs impact fish behavior in the wild? Yeah. So to be able to study this in the wild, we used slow release implants to be able to expose the Atlantic salmon to the cocaine and its metabolite. And to do these implants, what we did was dissolve the cocaine and its metabolite in a fatty carrier that we were able to inject into

the body cavity of the fish. And then when it was in there, in the body, then it slowly releases the pharmaceutical, in this case the cocaine and its metabolite over time. Oh, okay. So it's a slow release that's to represent what they would experience in the water. Exactly. It, allows us to give the salmon in this study a sustained exposure over time without it being a super high dose all at once that clears really quickly. How did you monitor their behavior? While we were giving the fish their implants with the cocaine and metabolite, we also implanted them with a small tag that allowed us to track where they are in the wild.

And then when we release them in Lake Veteran, there's an array of acoustic receivers that are like little microphones spread out in the environment. And the tag releases a ping in a really high frequency that these receivers can pick up on. And so later on, we can then go and collect our receivers and kind of triangulate the position of where the fish are in the wild without needing to follow them or track them manually ourselves. So once you release these fish that have been implanted with the slow-release cocaine, the metabolites, and a little sonic tag, what did you find? Yeah, so the fish that were exposed to the metabolite, they swam one and a half times further per week than controls by week five of the experiment. And then by week seven, they were swimming almost two times farther. Wow. Well, what went through your mind when you saw these results?

I first thought it was really interesting, especially because it was the metabolite that was having the most effect. And that's interesting to me as an ectosychologist because when we monitor for pollutants in the environments like pharmaceuticals and illicit drugs like cocaine, we tend not to look for the metabolites. It's just not part of active monitoring programs. And we don't do a lot of research to understand what impacts they might have on wild animals. animals. So it really kind of highlighted to me that metabolites are something that could be having an effect that we're really not looking at right now when we're trying to understand the impacts of pollution in the environment. Well, how concerning is this for the salmon with their exposure to both the cocaine and its metabolites? Basically, if you're swimming further in the natural

environment as a salmon, you basically have more potential opportunities to use different habitats. you could find different sources of food, but this could also expose you to more risk, poorer habitat, or potentially expose yourself to new predators. Oh, so this is not necessarily a bad news story. I mean, there could be benefits to this. I mean, if they could swim farther, that might give them a better chance at, say, finding a mate.

That's an open question that we would need more long-term studies that would extend beyond eight weeks and follow these salmon across their lifetime to understand the ultimate effects. Now, you were studying one drug at a time here. Yes. And how might that become more complicated? As you say, there are many different drugs in the water and how those drugs might interact with each other. It's almost a cocktail of contaminants that are present in wastewater effluence.

And some of them may act through a similar pathway or a mechanism of action. So you could expect, oh, maybe the effects would add together and be more worried. But then there might also be compounds that could have the opposite effects. So if cocaine's in upper, maybe there's also downers in the wastewater effluent. It's really hard to make those estimations of what those effects might be. This was at least a first step to understand what the effects of cocaine and its main metabolite. If these drugs and their compounds are coming through wastewater treatment plants, is there anything that can be done to the treatment plants themselves to try to filter them out?

Yes, there is. There are newer wastewater treatment technologies like ozone sterilization or activated charcoal filtration that are more designed to remove organic contaminant compounds like pharmaceuticals. How difficult would it be to do that? It's not necessarily how difficult it would be. It would be how expensive it would be. Should we be concerned about these chemicals and drugs appearing in our drinking while? Usually where drinking water is coming from is not the same location that wastewater is being discharged into. So there's going to be a geographical separation. But all of our drinking water is also treated again before we drink it. And it's held to a higher standard of what kind of levels of pollution are acceptable for drinking water than for the wastewater.

Oh, that's good news. Yes, that's a good news story. Dr. McCallum, thank you so much for your time. Thank you so much for having me. It was great to talk to you. Dr. Aaron McCollum is an associate professor of aquatic ecology at the Swedish University of Agricultural Sciences. Around four billion years ago, give or take a few hundred million years,

something remarkable happened on our planet. Until that point, there have been lots of physics and chemistry happening, but then, for the first time, something more like bio-earned, was taking place, as the first self-replicating molecules made their appearance and life came to be. But what exactly was that molecule? For some years now, the focus has been on RNA, a close cousin of DNA. That's the twisted, double-stranded molecule that acts like a recipe book with all the instructions to build the bits that make us, as well as other forms of life. Well, RNA is essentially just a single strand that can come in many sizes and shapes. And there's one very special type of RNA that not only stores

genetic information like DNA, but it can also serve as a kind of catalyst that may have kick-started the chemical reactions needed to make life work. But scientists have been trying to get these often large and unwieldly special RNA molecules to copy themselves without any luck, until now. A team of scientists at the University of Cambridge in the UK have identified one of these special RNA molecules that's small enough to form spontaneously, yet sophisticated enough to copy itself, and it might finally offer clues about how life began. Dr. Philip Hollager is a molecular biologist at Cambridge and led the team that published the findings. Dr. Holliger, welcome to Quarks and Quarks.

Hello. First of all, how would you explain to a non-scientist why these special RNA molecules are believed to have played a role in the origins of life? I think the big question in the origin of life is how chemistry kind of turns into biology. How do you transition from chemical reactions that, you know, when they happen, they happen always the same time in the same way. how you transition to something like biology, which has a heredity kind of that you can hand, where information gets passed on from generation to generation. And one attractive idea is that this transition involves, you know, the acquisition of such a molecular memory, and RNA is a very attractive molecular candidate for that, kind of to conceive to have formed spontaneously on the early Earth.

Well, tell me about the RNA molecule that you did work with. So we discovered this kind of seaving through a large pool of random RNA sequences, and we designed those sequences to be small because we were specifically looking for a small RNA molecule. And then we tried kind of to have it copy itself. And again, you know, we found that it could copy itself with really, you know, very respectable accuracy, making something in the order of, you know, five to six errors every hundred letters. So would you characterize the self-replicating RNA as alive? Or is it chemistry or is it biology? I would certainly not call it alive, but it maybe begins to enter this gray zone between chemistry and biology. The way we think about it is that initially there was just chemistry,

then molecules began to enter this sort of gray zone kind of of being not a light, but beginning to exhibit the type of, you know, behaviors which we, you know, think are sort of unique to biology, like the ability to grow, to make copies of themselves, to generate offspring, and also to evolve. So then what kind of conditions do you envision that would have made that process happen to go from the stage that you're at to the actual emergence of life on Earth? I wish I knew this in more detail. Unfortunately, we can't go back.

and see how it really happened. But one of the things that we found very interesting is that in the context of RNA, RNA can be quite a fragile molecule. RNA seems to be almost a perverse choice as a primordial genetic material, if you consider that the surface of the early Earth was likely rather unfriendly with boiling hot volcanic springs

and meteorites raining down, etc., etc. So we were wondering, if we could think of a environment where RNA would make more sense. And one of the things we came up with is actually water ice. So when water freezes, you know, if it contains also RNA and maybe some of the substrates that our RNA needs, the water freezes out first, but all the molecules that are not water get excluded from the growing ice crystals into a brine that surrounds the ice crystals, and that stays liquid at sub-zero temperatures. And the amazing thing is that in those channels and pockets in the ice, the RNA, our RNA stays active. And it's a little bit slower

than at higher temperatures, but it's a bit like the tortoise and the hair. It just keeps going and going and going and makes more RNA. So I think we like to think of ice as maybe a sort of protective environment for RNA, if you want the nursery, where RNA could sort of take its first baby steps of self-replication before eventually venturing out into the world. So what does this tell us about the ease or the difficulty of life taking hold? I'm not sure ice makes the origin of life more or less likely, but I would say this, if we think of our solar system as a fairly typical planetary system, it is worth noting that liquid water only occurs on one planet, while water ice is found on many of the celestial bodies. In fact, in the outer solar system, there are whole

celestial bodies like the moons, some of the moons of the gas giants who seem to be made mainly from water ice. So you're talking about Europa and around Jupiter and Enceladus around Saturn? Exactly, yeah. So you suggest that these would be a good place. to look for life, in other words. So suppose that happens. I mean, there are missions that are going to Europa to try to find out how much water it has and how thick the ice is and whatnot.

Suppose we do find life there. What does that say to you that it happened twice in our solar system? We, of course, then, would have to really study what is there to understand if life occurred twice independently or if there was some sort of cross-contamination, some cross-fertilization between a different body. bodies in the solar system. I mean, clearly, the most hospitable planet for life is Earth, but we do not know for sure that life actually originated on Earth. Maybe Earth kind of simply

got colonized, kind of, because it was an even nicer environment that were life originally originated. So in your mind, then, what's the scenario that started life on Earth? I think the way I'd like to think about it is that pre-bauty chemistry created. the building blocks of life and assembles them into these random RNA sequences. It's a little bit like the famous, you know, monkeys and typewriters kind of, you know, there's many, many chemical reactions going on, all of these typewriters synthesizing a sort of random string of RNA letters. And then some of these strings make sense. It's a little bit like typing random text by accident. Some of the text will make sense. And some of these strings will simply see.

say, copy yourself, make more of yourself. And those molecules will begin to, you know, make copies of themselves. And in doing so, they will make some mistakes. And then some of these new variants kind of will be better at making copy themselves. And that's how evolution gets started. And I think that gets a sort of snowball rolling towards life. Well, the question of how life began on Earth is one of the great unanswered questions in science. So how satisfying is it for you to have found this RNA molecule that might have been able to copy itself and potentially kickstart the emergence of life and help answer that question. Discovery in science obviously is always super exciting. And I think what makes it interesting

and rewarding is that discovery often comes from corners where you don't expect it. Personally, I was skeptical that this approach that we took in this case was going to work, but it was that much more exciting when it turned out to work. And the, these molecules kind of could ultimately then begin to show a capacity to make copies of themselves. And I think that gives us a first glimpse of how those processes could have occurred on the earlier Earth. And certainly for me, that is very exciting. And if it happened here, hopefully it can happen somewhere else. I think that is still a big question, kind of like, is the universe actually teeming with life

or are we alone? And we still do not know. And I think this depends on a lot of, a lot of contingencies. Dr. Holliger, thank you so much for your time. My pleasure. Dr. Philip Holliger is a molecular biologist at the University of Cambridge in the UK. You may have heard the saying, if you fall off a horse, get right back on. It's meant to motivate you to help you move on from a setback rather than giving into fear. But what if you actually fell off a horse? Should you get right back on? New research suggests that that might depend on how you are feeling, because it turns out, horses can literally smell your fear and react in kind. Dr. Plotin-Jardot led this study. She's a researcher in equine behavior

and welfare at the French Institute for Horse and Riding in Tour France. Hello and welcome to our program. Hello. First of all, tell me about the historical relationship between humans and horses. How far back does that go? Horses have been domesticated by humans several thousand years ago, so that's quite far back, although not as far back as dogs or other domestic mammals. And through this process and over time, we have evolved together, so this can be an explanation why horses can smell how we feel. And what else can they detect in humans besides their smell?

So previous research that we've done in the lab has shown. that they can detect our emotions from our faces and voices to facial expressions and the tone of the voice. But they can also do a lot of things like they recognize different faces of people. So yeah, they can recognize people from their faces. They can also understand some interactions between humans and understand to some extent what we are looking at, they are really good at understanding us. So we have a real connection with horses, then, like an emotional connection with them?

Yes, I would say that. So in terms of scent, what is it that the horses are picking up if we're, say, fearful? That's something we don't know yet. We know that from our sweat, they pick something up and they can detect it and get emotional information. but the chemistry of this is not yet known. So some studies are being done at the moment to understand this more. But we have some ideas about this because there have been some studies in humans that show some particular molecules that are known to carry some emotional information

or at least to be different in our sweat when we feel different emotions. But we don't know yet how it works with horses. Walk me through your experiment then. How did you study whether horses can pick up our fear or not? First, we collected some sweat samples from people. So we had a lot of nice volunteer people come to the lab. And they watched two different movies. So on one day, they would watch a scary movie.

It was from a horror film. And on another day, nice video clips, for example, some musicals or some jokes. And each time they were wearing cotton pads under their armpits. That's how we collected their sweats while they were feeling some emotions. So fear on one side and joy on the other side. Then with these cotton pads, we had horses smell them. So we put them in front of the horses' noses.

And while the horses were smelling these sweat samples, We performed behavioral tests, and we also measured some physiological variables like their heart weight and saliva cortisol. Wow. So you didn't have the people themselves in front of the horses. It was just their scent, just their sweat. Yes, that's it. We wanted to isolate this. As a first step, it's easier to just collect the sweat than have the people in front of the horses. also because this allowed us to be sure that the right emotions had been felt by the people. So if someone had not been afraid of the horror movie, we would not use their samples. So how did the horses react to these different sense? They reacted really differently according to the emotion.

We had three groups of horses. One group was smelling the sweat samples from the fearful. condition, one group from the joyful condition, and one group was smelling control others, so that would be fresh cotton pads. And what we saw was that the horses smelling the others from the fearful condition were behaving differently from the others. And actually, they showed that they were themselves more afraid. The horses were more afraid.

Yes. So there were two situations in which this was the most clear. The first situation was a suddenness test. So we would open an umbrella close to the horse's head while they were eating something. So that's startling, right? It would be for you too, I guess. And so when we do this close to horses' head, they are startled. They have a little jump scare.

But for the horses that were smelling sweat samples from humans feeling fear, they jumped further or higher and also their heart weight rent higher. So this showed that they were more afraid of this. Well, what does this say about our emotional connection to horses? Well, I think this says that this connection is really deep and it can come from visual exchanges, but also, from what horses hear from us, but also from our odors. And I think what's really interesting is that we can control what we say, how we say it, or our facial expressions to some extent. But for smells, we can't really control what we are producing.

So our smell is really something that horses can pick upon maybe before us, Like, sometimes they can detect that we are afraid of something when we are not aware ourselves. It's almost like we're passing our emotions on to them, because if they become more startled, more easily startled, we're giving them our fear in a way. Yes, exactly. That's what we call emotional contagion. So that's when an emotion goes from one individual to another. and we are only starting to see that this can happen between different species. Be calm and confident. Dr. Jardat, thank you so much for your time. Thank you for having me.

Dr. Plotin Jardat is a researcher in equine behavior and welfare at the French Institute for Horse and Riding in Tour France. As we mentioned last week, we're gearing up for another listener question show. Our inbox is already filling up with some pretty fun brain teasers but we want more. So please continue to send us your science questions and we'll get to work finding you some answers. So email us at quirks at cbc.ca.

And that's it for Quirks and Quarks this week. If you'd like to get in touch with us, our email is Quirx at cbc.ca. Our webpage is cbc.ca. slash quirks, where you can check out our past episodes and find more information on the research we covered in the show. You can also follow our podcast, get us on SiriusXM, or download the CBC Listen app. It's free from the App Store or Google Play.

Works and Quarks is produced by Sonia Biting, Rosie Fernandez, and Dan Falk. Our intern is Sarah Hamilton. Special thanks to CBC Radio Archives, Patrick Mooney, Ross Tully, and Zoe Barraclough. Our acting senior producer is Amanda Bukowitz. I'm Bob McDonald. Thanks for listening. CBC Podcasts, go to cBC.ca slash podcasts.