Host: Shamini Bundell
Welcome back to the Nature Podcast. This week, how diversity helps prevent food supply shocks.
Host: Nick Petrić Howe
And the rise and risks of computational social science. I’m Nick Petrić Howe.
Host: Shamini Bundell
And I’m Shamini Bundell.
[Jingle]
Host: Shamini Bundell
In an era of increasing instability, from the COVID-19 pandemic to climate change, food insecurity is a growing problem. This week, a new paper out in Nature models how cities can adapt their supply chains to better keep food on their citizens’ plates. Reporter Ali Jennings talked to author Alfonso Mejia to find out more.
Interviewer: Ali Jennings
This paper is about protecting cities from food supply shocks, and what is a food supply shock?
Interviewee: Alfonso Mejia
So, cities receive influx of foods for different food types, and when a city experiences a sudden decrease in this influx we call that a food shock.
Interviewer: Ali Jennings
And what can cause that?
Interviewee: Alfonso Mejia
Yeah, so there are many causes for these food supply shocks. One main cause is climate change, climate reliability. As we experience more droughts it can have a strong effect on crop production. Geopolitical instabilities can be a problem. I mean, the more recent one perhaps is more of a demand-induced shocked that we experienced with COVID-19, where people are not going out to eat at restaurants. That decreases demand for food and that can also cause a shock.
Interviewer: Ali Jennings
What is the effect of food shocks in a city?
Interviewee: Alfonso Mejia
They can have an influence on food prices. If the food prices go a little bit up then low-income or vulnerable populations in a city can be impacted. So, if you look at low-income, underserved communities, like Blacks and Hispanics in the United States, you see these percentages of food-insecure people going up in cities of the US.
Interviewer: Ali Jennings
So, cities have this problem. They can be vulnerable to food shocks. What did you set out to do in your study?
Interviewee: Alfonso Mejia
We wanted to look at whether there are lessons from nature that we can learn in the way we design our food systems and in the way we design our cities. So, specifically we wanted to look at whether supply chain diversity can help cities cope with food shocks. We know from ecosystems that greater biodiversity can be beneficial in terms of protecting ecosystems against external shocks. So, we wanted to check that hypothesis on our food systems in cities.
Interviewer: Ali Jennings
And how did you do it? What were your methods?
Interviewee: Alfonso Mejia
Yeah, so for us to do that, we worked with a large dataset describing the food supplies from regions in the US to cities in the US, so we had around maybe 300 cities in our data. And then we developed a measure of supply chain diversity where a city has a more diverse food supply when that city links to trading partners that are more different.
Interviewer: Ali Jennings
Could you sum up what you found?
Interviewee: Alfonso Mejia
Yeah, so the main finding of our research was that the probability of a city experiencing a food shock tends to decline with increasing supply chain diversity. So, a city that sources food from different climate regions in the US and perhaps using different infrastructure systems to bring that food to the city.
Interviewer: Ali Jennings
You based your model on data from only American cities. Can you generalise to other cities across the world?
Interviewee: Alfonso Mejia
The United States is a large country. There’s a lot of variability in terms of city size, city demographic, characteristics. We think that there is enough variability inside the US that our results can be potentially used for cities elsewhere. There are these universal sort of features of cities that maybe useful to generalise our results further.
Interviewer: Ali Jennings
So, your model predicts that increasing diversity makes a city less likely to suffer from food shocks, but what does that look like in practice?
Interviewee: Alfonso Mejia
This will obviously require a lot of planning and coordination. Cities themselves can create maybe the policies, the incentives, the environment for companies, for retailers, wholesalers to create operations that are more diverse. The government, at least in the United States, is heavily involved the food supply chain. There are food programmes for low-income groups, for schoolchildren, for the elderly, so there might be a way there for cities to coordinate with government at different levels to think and to develop and implement practices that promote supply chain diversity.
Interviewer: Ali Jennings
And what are the trade-offs for making your food supply more diverse?
Interviewee: Alfonso Mejia
Obviously it adds complexity to the supply chain. It also potentially exposes you to disruptions in other places, so as we diversify supply chains to reduce food shocks, we also have to be sort of aware that there are limits to diversity, that there can be levels of complexity that become attainable, but I think we are now at a moment where we are more aware of these level threats. Recently with the pandemic we all experienced a systemic shock, and I think cities are perhaps thinking more about resilience these days and maybe it’s a good opportunity now for them also to think about resilience to other types of shocks and in particular to food supplies.
Host: Shamini Bundell
That was Alfonso Mejia from Pennsylvania State University in the US. For more on how to mitigate food shocks, be sure to check out Alfonso’s paper. We’ll put a link to it in the show notes.
Host: Nick Petrić Howe
Coming up, we’ll be finding out about an up and coming field of research that’s using big data to tackle social science questions. Right now, though, it’s time for the Research Highlights, read by Dan Fox.
[Jingle]
Dan Fox
A precisely printed 3D sheet no thicker than a human hair has withstood the impact of miniscule bullets crashing into it. Researchers 3D printed the sheet from a flexible polymer that was then heated to 900 degrees, leaving a carbon lattice resembling a honeycomb. They then tested its ability to dissipate force by firing tiny spheres of the material at up to twice the speed of sound. A lower speeds, the micrometre-scale spheres simply bounced off the lattice. But at higher speeds, they gauged out neat craters, crushing the lattice underneath and remaining lodged in the material. By adjusting a model used to describe meteor craters, the team were able to predict crater formation in their materials. They hope that this model can inform other lightweight, blast-proof designs. Judge the impact of that research for yourself in Nature Materials.
[Jingle]
Dan Fox
Can you remember the running order of the podcast so far? Well, to recall a past experience, we need to piece together when specific events happened and in what order. Now, scientists have confirmed that humans have ‘time neurons’ in their brains to encode this information. Rodents have long been known to have time-keeping cells in the hippocampus. To look for similar cells in humans, researchers recorded electrical activity in the hippocampal neurons of volunteers as they memorised a series of images and were periodically quizzed on their order. The researchers pinpointed a subset of neurons that fired as the participants completed their activities. These neurons showed similar patterns of activity during gaps when participants stared at a blank screen, suggesting that they were encoding time rather than other information about the task. Take the time to read and recall that research in full in the Journal of Neuroscience.
[Jingle]
Interviewer: Nick Petrić Howe
This week in Nature, we have a special issue all about computational social science. To drill into the details, joining me today are Federico Levi and Mary Elizabeth Sutherland, who have been instrumental to this special issue. Hello to you both. Would you like to introduce yourselves?
Interviewee: Mary Elizabeth Sutherland
Hello, it’s a pleasure to be here. My name is Mary Elizabeth Sutherland and I am a senior editor at Nature handling manuscripts in the behavioural sciences, which means that I read papers that have to do with humans, what they think, feel and do.
Interviewee: Federico Levi
Hi, I’m Federico Levi. It’s great to be here. I’m a senior editor in the physical sciences team at Nature, so I’m a physicist, and I’m responsible for handling submissions from fundamental physics, computer science and applied mathematics.
Interviewer: Nick Petrić Howe
So, you’re both from quite disparate fields and that may, in a way, speak a little to this topic, but before we get into that, I thought a good place to start would be, well, what is computational social science?
Interviewee: Mary Elizabeth Sutherland
So, computational social science is really represented by Fede and myself. It’s a computational approach, so a big-data approach, understanding things using computational techniques, modelling, analyses that are only possible with a lot of computational power, applied to questions in the social sciences.
Interviewer: Nick Petrić Howe
Big data tackling some these social science issues, what would be a good example of something that would fit into computational social science?
Interviewee: Federico Levi
I mean, there is quite an array of things but you could imagine something going from understanding how and where people move on the basis of their GPS traces on their mobile phones, and then trying to understand what are the drivers of their behaviour, in terms of how people react to some event, whether there is a preference to use a certain transportation means.
Interviewee: Mary Elizabeth Sutherland
I think a good example that sort of goes to the mobility example, Fede, is inequality as well, right? So, instead of just having traces of where people move, you can start to understand social structure and what people with different demographics, different social economic status, how they move differently, which can then inform you about the description of inequality in different societies and then also give you ideas of how that can be changed. What are the things that make the effects of this inequality less bad, I want to say?
Interviewer: Nick Petrić Howe
And what, would you say, has prompted this special issue in Nature?
Interviewee: Mary Elizabeth Sutherland
There are two main things that I’d like to highlight. One is that the amount of data and the granularity of the data that we are now generating due to smartphones and smart devices like watches has given us an unprecedented range of data to do this computational social science, and so that has really caused a rapid expansion in the field, so it seemed like a good time to stop and take stock of what the challenges are, how we can work together to make sure that computational social science does really good, strong and ethical work. So, that’s one main motivator. Another motivator is to showcase that Nature does consider work in the social sciences because it’s not a field that we have been known for considering in the past.
Interviewer: Nick Petrić Howe
This sounds like there’s a lot of potential in this field. There’s a lot of interesting questions that this could potentially help with, but you’ve both said there are several challenges involved with this sort of work. Could you tell me a little bit more about what the challenges are in this field?
Interviewee: Federico Levi
At some level, a very practical challenge is finding the data that would be necessary to carry out these studies because the data is there but it doesn’t mean it’s actually very easy to access. And then the sort of data that is available, the quality of this data and the bias that is inherent in this data will then substantially impact the sorts of studies that can be done. The most obvious example is it’s most likely fairly wealthy people from the western side of the world that generate the sort of data that these studies can be based upon, and this of course limits the reach of the sorts of studies that one can do and this is something to be kept in mind.
Interviewer: Nick Petrić Howe
So, I guess when we’re talking about big data and how people are interacting with smart devices and things, one thing that comes to mind are sort of ethical issues, how this data is being used. Is that something that’s covered in the special issue?
Interviewee: Mary Elizabeth Sutherland
Yes, so when you have a lot of granular data about somebody – again, just think about mobility data – even if it’s anonymised, let’s say you see the location of the phone and in the evening at night it’s always in the same place. It’s very easy for you to then assume that that’s where that person sleeps and lives. So, you can get sensitive information even from anonymised datasets. So, what’s really key is thinking about what questions you can answer without needing the granularity of data that could potentially give away sensitive data.
Interviewee: Federico Levi
And I think something else is also a matter of like where are these studies going. Is this research truly going to improve people’s lives or improve society or is this going to lay the basis for using these insights to manipulate or carry out surveillance on people? This doesn’t mean that this research shouldn’t be carried out, it just means that people that carry it out need to be aware of these challenges and act accordingly.
Interviewer: Nick Petrić Howe
So, there’s a range of challenges within this field then. Do we have a sense of how these barriers, these challenges may be overcome?
Interviewee: Mary Elizabeth Sutherland
Yes, the papers in the special issue go into a lot of detail about ways in which some of these challenges can be overcome. One of the challenges is understanding the influence of algorithms on behaviour. So, just think of anything that you’ve done recently. Let’s say you wanted to go out to a restaurant and you didn’t know which one to go to. You might have searched Google for ‘Japanese food’ or you might have looked at Yelp for a recommendation. Both of these are examples of how an algorithm can influence your behaviour. You are choosing your restaurant based on what comes up in Google or how many stars something has gotten on Yelp. So, one of the papers is called ‘Measuring algorithmically infused societies’, and the idea in this paper is how do we deal with the influence of algorithms? And the solution that they propose is one that is intuitively simple but very important, which is to acknowledge that most of the behaviour we see people do has been influenced by algorithms. So, they make an argument for how we can acknowledge this influence of algorithms and then how we can build better theories that take into account that humans are now not just interacting with other humans but that our day-to-day life involves algorithms.
Interviewee: Federico Levi
So, another barrier that computational social science faces is the fact that this sort of research requires very interdisciplinary teams and, as happens often, very different scientists speak very different languages and approach problems in very different ways. I think that in this case, given that we’re trying to bridge the divide between social sciences and natural or computational sciences, the distance is even wider. So, it’s really an effort to try to build productive and fruitful collaborations and we have one of the articles that suggests a way to do so by essentially analysing the way in which research is carried out in these disciplines and say, ‘What are the building blocks of the approaches that people use in the natural/computational or social sciences,’ and basically identify what are their strengths and key features and suggests that productive computational social science can come out when not only teams are interdisciplinary but even the research approaches really incorporates these sort of strengths that come from the approaches of the social or the computational sciences.
Interviewer: Nick Petrić Howe
And one other thing I was wondering about as well, I think this pandemic has for a lot of people put into sharp focus the power of big data, looking at epidemiology, case numbers and that sort of thing. Do we have any sort of insights into sort of disease modelling and the influence of the pandemic from this special issue?
Interviewee: Federico Levi
Yes, we actually do. There is an entire article that touches upon the fundamental interplay between human behaviour, societal structures and the diffusion of infectious diseases, and the authors argue that it’s actually very hard to carry out thoughtful epidemiology and epidemic modelling or forecasts without really sitting down to understand the way people behave in the face of something quite as scary as an epidemic or a pandemic. So, the pandemic made a lot of data available but once again, as tempting as it is to just draw from it and draw rapid, sweeping conclusions on case forecasts, the reality is much more nuanced than that and good epidemiology and particularly the epidemiology of the future is probably going to need to be built on fundamental social understanding that probably will have to come from social sciences.
Interviewer: Nick Petrić Howe
And speaking of the future, do you both have any predictions of what we might see from this field going forward?
Interviewee: Mary Elizabeth Sutherland
I think we’re going to see a lot of exciting things. I think we’re going to see a lot more of research that pertains to more and more of the population. So, as more and more people get cell phones and have access to the internet and the more that our digital infrastructure expands, the more that we will be able to get data from a more representative swathe of the human population and the more we will be able to understand how we are similar and how we are different.
Interviewer: Nick Petrić Howe
So, there’s certainly a lot to keep an eye on in the future, but I think that’s all we’ve got time for, for now. Listeners, I’ll put a link in the show notes which will take you to a landing page for the special issue where you can find lots of information and the articles that Fede and Mary Elizabeth spoke about. But for now, Federico and Mary Elizabeth, thank you for joining me.
Interviewee: Federico Levi
Thank you so much.
Interviewee: Mary Elizabeth Sutherland
Thank you for having us.
Host: Shamini Bundell
Finally on the show, it’s time for the Briefing chat, where we take a look at some of the highlights from the Nature Briefing. Nick, what have you been reading about this week?
Host: Nick Petrić Howe
I’ve been reading about viruses, and this is a feature in Nature this week and, well, I think we’ve gotten used to the fact that a virus can have a big impact on the world – one particular virus – but that’s also true of the rest of viruses too.
Host: Shamini Bundell
And the rest of the viruses haven’t really been on our radar much this year. They’ve been sort of somewhat overshadowed by their celebrity cousin, but which particular viruses is this feature focusing on?
Host: Nick Petrić Howe
Well, all of them really. The feature really focuses on the fact that we’re just now starting to understand what a huge diversity and number of viruses there are out there, and that’s because in the past ten years or so, there’s been an increase in sequencing technology and things that have allowed us to get a better understanding of the genomes of viruses. Because the trouble with viruses in the past is you had to culture them in order to study them, but now with sequencing technology, that’s not always the case because you can send a scattergun approach and just get samples of the environment and sequence all the genomes in there and get an idea of how many viruses there are there, and it seems like there are billions and billions and billions and billions of them.
Host: Shamini Bundell
And what kind of environments are you talking about? Are we talking about just disease-causing viruses in humans, that we’ve got more than we think?
Host: Nick Petrić Howe
Well, we do have more than we think but there’s also like a range of viruses in the environment. Like just about anywhere on Earth there are viruses, so in the deep oceans, in the human gut. So, there’s one study that’s mentioned here, for example, where they looked inside the human gut and found 140,000 different viruses that infect bacteria alone. But the thing is, there are only about 9,000 or so that have been named because, again, we were relying on this ability to culture them in the past. So, now that we’ve got these sorts of sequencing technologies, we’re able to sort of get an idea of a greater range of them and the numbers of named viruses is dramatically increasing year on year. In 2020 alone, 1044 species were added to the list and, when there’s only 9,000 in total, that’s quite a big jump.
Host: Shamini Bundell
And presumably, so viruses can’t live on their own – they’re sort of parasitic depending on other organisms such as bacteria – so, in the olden days when you say they would have to culture the viruses, presumably in order to find out the virus of any particular bacteria or like weird sea creature or animal, you’d also have to be able to then grow that animal or have cells of that animal.
Host: Nick Petrić Howe
Yeah, so, I mean, that’s the difficulty. It’s not easy to culture these things in a lab. But with genomes, they’re just there in the environment. You can go through and you can sort of find them and then start to understand some of the diversity and because of this, we’re starting to get an idea of viral evolution as well. And from what we understand of it now, it seems like there is good evidence that there wasn’t one single thing that became a virus and then all viruses came from it. Actually, viruses originated multiple times.
Host: Shamini Bundell
Oh, wow. Okay, so did people know this before? I guess I just always assumed that there was one first virus and that’s the sort of family tree.
Host: Nick Petrić Howe
I think theories have been proposed before but we just didn’t have enough information to really figure it out. Now we’ve got a lot of genomes, we’re starting to understand like how the different parts of the phylogenetic tree, like the family tree of viruses for lack of a better term, fits together. But the trouble with viruses as well is they shuttle genes around between them quite a lot and things, so it’s quite hard to work out where they all fit together, and some viruses mutate very quickly so they’re very diverse so it’s hard to work out what gene belongs to what, but we’re starting to pick this out now. So, you may be familiar with the sort of domains of life. Well, they’ve done a similar thing for viruses but they’ve actually gone with a much cooler name in my opinion, which is they’ve named them ‘realms’. And so, there’s about 25 proposed realms of viruses of different kinds of viruses and how they fit together, so things like whether they have RNA or they have DNA and things like that.
Host: Shamini Bundell
It’s so fascinating and even the fact that viruses aren’t technically considered a sort of living organism but they’re these amazing constructions that can evolve multiple times, can evolve and change and have relationships and phylogenetic trees just like other organisms. Amazing.
Host: Nick Petrić Howe
Yeah, and also, because there are so many of them and they are so diverse, there’s an increasing understanding that they’re really affecting the environment as well. So, for example, many viruses will cause bacteria to explode and release all their contents and that can actually allow a lot of carbon to end up in the atmosphere because all the bacteria gets exploded out, then that gets eaten by other microorganisms and then that can get excreted into the atmosphere. But because there’s so many of them, it can actually have a quite large impact on it and it could also sequester it. So, for example, if they do that in the ocean, then those bits of bacteria might sink to the bottom of the ocean and sequester a lot of carbon.
Host: Shamini Bundell
So, these little things, I mean, we understand now that they do have big impacts on the world, but they may have impacts far beyond just what they’re having on humans.
Host: Nick Petrić Howe
Yeah, so not just diseases but full on food webs and geography and even climate change.
Host: Shamini Bundell
Are you saying that viruses are really important for climate change?
Host: Nick Petrić Howe
Yeah, I mean, they certainly seem to be having a big impact on a lot of things, and now the question is: which ones do we study?
Host: Shamini Bundell
Well, it’s nice to have a good news virology story for once. That makes a nice change. And I have also some good news from the world of certainly diseases if not viruses in this case. I’ve been reading an article in Science about the elimination of malaria in China.
Host: Nick Petrić Howe
Wow, so it’s completely eliminated in China? That’s very impressive. I mean, how much of a problem was it there before?
Host: Shamini Bundell
So, China has been working on this for decades. So, this maybe started back in the 50s, and in the 40s and 50s they had 30 million cases a year in China and 300,000 deaths, so it was a huge problem and they have spent the whole time since then getting those numbers down and finally they haven’t had a case that’s originated from within the country for the last four years, and the World Health Organization sent some people over and they announced that yes, China is joining the list of countries that’s been announced to have officially eradicated malaria.
Host: Nick Petrić Howe
Well, going from 30 million to zero sounds very, very impressive. How did they actually go about it?
Host: Shamini Bundell
So, quite a lot of different strategies and also quite a lot of innovation in ways to fight malaria. So, for example, back in the 80s, China had really pioneered the use of mosquito nets treated with insecticides. They also had a big programme to try and identify new malaria drugs in the 60s, and there was a particular very successful drug that came out of this – artemisinin – which was discovered by pharmaceutical chemist Tu Youyou who won the Nobel Prize for it in 2015 at age 84, so this has been a lot of work and a lot of contributions from a lot of different people.
Host: Nick Petrić Howe
And you said as well about sometimes the cases coming from elsewhere. Is there a worry that it might return in China?
Host: Shamini Bundell
Well, yeah, so, China borders countries that malaria is still endemic in. It could easily travel over. So, part of the WHO certification is showing how you’re going to prevent it coming back and kind of restarting again. You can’t just get rid of it once unfortunately. It’s a bit of an ongoing battle.
Host: Nick Petrić Howe
So a lot of strategies then have been employed. Could a similar sort of thing be done in other countries where malaria is still present?
Host: Shamini Bundell
Yeah, so, at the moment, the biggest problem is for African countries, so Africa has most of the existing cases at the moment. And hopefully the methods and techniques and drugs developed in China and put to use, because China is an immensely populous country, it’s the most populous country to have been so far declared malaria-free, so hopefully some of that can help other countries as well.
Host: Nick Petrić Howe
Well, fascinating stuff, Shamini, and hopefully this can be somewhat of a roadmap for other countries. I think that’s everything for the Briefing chat this week. Listeners, if you’re interested in more stories like this then why don’t you check out the Nature Briefing. We’ll pop a link of where to sign up, along with the stories we discussed, in the show notes.
Host: Shamini Bundell
That’s all for this week’s Nature Podcast but we do have a whole host of new video content up for you. We’ve got short films about diabetes, an artificial pancreas, and the physics of how leaky pipes move water around, so check out the links to those.
Host: Nick Petrić Howe
As always, you can reach out to us on Twitter – we’re @NaturePodcast – or send us an email – we’re podcast@nature.com. And if you like what you hear, why not leave us a review? Do that wherever you get your podcasts. I’m Nick Petrić Howe.
Host: Shamini Bundell
And I’m Shamini Bundell. Thanks for listening.