Meet the panel: going vegan and cancer AI

07 January 2020

Interview with 

Tim Revell, New Scientist; Ella Gilbert, British Antarctic Survey; Jess Wade, Imperial College London; Hannah Thompson, Cambridge Cancer Genomics

COW

Cow in a farm

Share

With us this week are Tim Revell, tech journalist for New Scientist magazine; climate scientist Ella Gilbert from the British Antarctic Survey; Jess Wade, physicist and award-winning science communicator; and Hannah Thompson, geneticist and Chief Product Officer for Cambridge Cancer Genomics. Chris Smith sat down to ask them about going vegan and AI that can detect cancer...

Chris - Now with us this week, our talented tech head, Tim Revell, he's from New Scientist. Great to have you back, Tim.

Tim - Yeah. Great to be here. Thanks!

Chris - And what have you had on your radar at New Scientists lately?

Tim - Well, we've been testing out the idea that going vegan improves your carbon footprint, reduces the amount of CO2 you are responsible for. And some of us did an experiment in the office where we recorded everything we ate laboriously for a week. And then the following week everyone went vegan and recorded everything they ate two and then a specialist analyzed everything we'd eaten and looked at how our carbon footprint was affected.

Chris - How does that actually work? Why would going vegan translate into a superior carbon footprint for a person who's normally a meat eater?

Tim - So there are certain foods that, when you look at how they are produced, how much food for example goes into making them. So beef for example, you have to feed a cow and then ultimately you kill that cow to eat the meat. Things like milk. You also have to keep a cow alive. All of those correspond to energy that goes into producing that food and some foods are more heavy in terms of the energy you need than others.

Chris - And how big was the difference?

Tim - It was a really large difference. So those people who were meat eaters who went vegan, their carbon footprint was reduced by around 70%. There were also people who are vegetarian already, like myself, who then tried to go vegan for a week to see whether it made much of a difference. And for those people it halved their carbon footprint too. So it does make quite a big difference.

Chris - Do you think it's sustainable? Do you think people will stick with it? Because it is quite hard if you're not used to eating like that to embrace a healthy vegan lifestyle, isn't it?

Tim - Yeah, I myself find it really difficult. I find it really easy to go vegetarian. I find it really hard to go vegan for a week. I found like you have to become an expert, whereas you can sort of be a casual vegetarian. But I think what we found is that by reducing things a little bit, you still make a difference. So you don't have to go to the extreme of becoming a hundred percent vegan to reduce your carbon footprint, if you're interested.

Chris - Thanks Tim. So any tech questions and perhaps even vegan questions, you can put those to Tim. Well, it's fair to say Ella Gilbert, who's also with us from British Antarctic Survey and who's going to talk about climate, you're vegan, is that right?

Ella - That is right, yeah.

Chris - But you were vegan before it was cool to be vegan. You were vegan all along! So why are you vegan?

Ella - For the environmental impact predominantly? I also find that it really helps my health. So I'm a boxer in my spare time and I find it a lot easier to stay in my weight category if I'm vegan. But I mean the main motivation was the environmental impact.

Chris - Of course, the environment's been in the news a lot, hasn't it? I mean not only because we've seen Paris and the implementation of the Paris agreement, but we've also got what's going on, what has been going on for a while now in Australia with these fairly dramatic fires.

Ella - Of course. I mean climate change is all over the news and I don't think it's going away anytime soon. And these sorts of dramatic events like we're seeing in Australia, they're happening more and more frequently. So they're not going away.

Chris - When you are working at British Antarctic Survey, what actually do study?

Ella - Predominantly I'm looking understandably at Antarctica. I do climate modelling a lot of the time, but trying to understand how melting occurs in Antarctica. What specifically is causing that melting. And I've got an atmospheric physics background. So it's, it's usually weather.

Chris - And boxing! So punching above her weight on climate science. That's Ella Gilbert. Also with us is Jess Wade. She is a physicist and someone who's interested in metamaterials. So that's the interface between physics and chemistry. She's Imperial college. She's also with us. Hello Jess. Welcome back. And what have you had your eye on?

Jess - I was super interested in this story that came out a couple of weeks ago about Google and DeepMind and a bunch of universities around the world looking at improving our efficiency of detecting breast cancer in people who've had mammograms, training, huge algorithms and data sets to be able to identify where within those images people were going to suffer from breast cancer.

Jess - So if you do a bunch of mammograms, I think they studied 28,000 ,to try and tell whether people had breast cancer and to be able to do it more efficiently. And they managed to do it, they managed to develop an algorithm that could do it as efficiently as doctors.

Chris - So just to be clear, this is showing a computer the same images that normally a radiologist, an x-ray doctor would look at and training the computer what to look for.

Jess - Exactly that. And something that we do in the UK is we have kind of double readings. So it will go to a first radiologist and then to a second radiologist. And if there's a discrepancy between the two, you'll call that person back in and do some more checking. And what this algorithm and AI tool can do is really increase that efficiency so that you don't have to have more and more of these expert radiologists. You make them have to do less work. And we have a big challenge in recruiting people with those skills. So it'd be really, really great if we had a computer that could do that for us.

Chris - Thanks Jess. Also with us is Hannah Thompson who actually is interested in genetics and genomics, you actually work at Cambridge Cancer Genomics. What's that?

Hannah - Yeah. Hi Chris. It's a startup. We're interested in getting patients the right treatment at the right time. So we actually use a bit of this aforementioned AI and we help to help the doctors make a decision about what treatment that patient should be on.

Chris - And is that based on looking at the genetics of a problem a person has or looking at their genetics and making predictions about problem they might have? How's it work?

Hannah - Yeah, so right now once a patient has been diagnosed with cancer, we take a blood sample and you can actually find DNA that relates to their cancer in their blood, which is really cool. And what we do is we look at those mutations that are there and see what the most applicable treatment is at that time. Now in the future we would really like to do what you just mentioned, predict everything that could happen and say from this single blood sample, we would love to have you on this treatment for four weeks instead of six months and then switch onto this one that's going to work even better.

Chris - How's that sound to you, Jess? Sounds great!

Chris - But I like your algorithm as much.

Jess - Yeah, I really liked the idea of it. And I come from a family of NHS employees, so like anything that can reduce the burden on the NHS.

Chris - Sounds good to me.

Comments

Add a comment