Transcript
RICHARD HOLLIMAN
I’m joined now by Claire Turner, who works in the School of Life, Health and Chemical Sciences at The Open University. We’re here to discuss how science progresses.
So Claire, what is your current topic of enquiry?
CLAIRE TURNER
Hello Rick, it’s very nice to talk to you today. My particular interest is in breath analysis and the use of – the odour of an individual to try and find out whether they have a particular disease or the state of their health.
RICHARD HOLLIMAN
Okay. That’s fascinating. So, what would you see as the kind of key characteristics of a successful scientist?
CLAIRE TURNER
Well, for me, I think there are two fundamental characteristics that every scientist should have. The first one is curiosity. If you really aren’t interested in the world around you, and some of these questions, you’re never going to make a scientist because you’re not going to really want to find out the truth, for science is about truth.
If you’re not curious about the truth and want to know about it you’re not really going to be able to work towards that. So that’s number one.
The second one, which is very much a part of the first, is scepticism. I think you can make mistakes if you are not sceptical, because if you have – most of us think – have an opinion about things. If you have an opinion about something which you develop into a hypothesis you will seek information which will be supporting your hypothesis and you won’t seek information which will refute your hypothesis.
You need to be able to seek information, which will just tell you about stuff around your hypothesis, and then you can decide whether that information and that data you’ve obtained supports or refutes it. So you must be sceptical. And I think that’s absolutely fundamental.
RICHARD HOLLIMAN
You’ve talked a bit about there about the kind of seeking truth if you like and one of the things I’m curious about is the notion of kind of agreed knowledge.
So what would be the kind of agreed knowledge in your area? What would be scientific truth if you like in your areas – established truth?
CLAIRE TURNER
My discipline is relatively young. It’s about 20 years old, but the idea for it came back from the Ancient Greeks when they kind of knew that if people had a particular smell coming off them, or off their breath, they had a particular condition such as tuberculosis or diabetes.
So they knew there was something wrong with them, couldn’t do anything about it, but they knew there was something wrong.
But it wasn’t until probably the late 1970s or around then that people started thinking that perhaps they could actually measure the composition of breath.
And Linus Pauling, a great chemist, actually used gas chromatography to analyse people’s breath, back in that time. But only in the last 20 years have people started thinking ‘Actually, they could use this to diagnose disease, it’ll be a lot easier.’
The problem is, although we do know that different diseases do give rise to different breath profiles or different odour profiles, lots of different groups have all been working independently from each other and there has been no standardised way of either taking a breath sample or sample of fluid from the patient, or actually the equipment used to – to analyse this has been different. Therefore, comparing what the different groups are doing might be the same disease, but they get different results.
And this has really been very difficult to try and therefore say, ‘Right, this disease has this exact profile, and this disease has this exact profile.’
So that’s really where the state-of-the-art is at the moment. We know that there different profiles, but it’s trying to nail those down because of the different groups working on it.
RICHARD HOLLIMAN
That’s really fascinating because one of my questions was going to be about how you see that kind – how you see those kind of established truths becoming established if you like.
So I guess I’d modify it slightly for the answer you’ve just given me and say, how do you set standards in your area, or how could standards be set?
CLAIRE TURNER
That’s an extremely good question and one which the field is currently looking at doing. So there are a couple of groups being set up at the moment which are trying to identify how to produce a kind of standardised breath sample, if you like, which can then be used in a way that’s agreed across all the people working on this and then used – used to actually validate the instruments and then we can go back and start looking at the existing data.
I think another issue is that studies generally tend to done on small numbers of people and the reason for that is it’s very, very expensive to do studies like this, to recruit patients, to actually get people in, to get scientists in, to get the instrument in where the patient is.
To do all that is hugely, hugely expensive and getting funding to do this for 50 or 100 patients; not too difficult, but because this is a new field if you like, and because there is such variability in individual people and well – biological systems are by their very nature very variable – you get variability just between people, but also within individuals you have huge variation over the day, and all sorts of things can cause huge variations, and as a result of that in order to make real headway in this, first you’ve got to standardise stuff but secondly, you really have to measure large numbers of individuals in a population.
So the only way you’re actually going to do this is to get lots of different groups working together, but they’re going to need to use the same methodologies.
RICHARD HOLLIMAN
So collaboration, standards established, and a larger data set.
CLAIRE TURNER
Absolutely and, of course, each set of samples you get on a particular condition on a particular group of people if it’s standardised that adds to the body of knowledge and gives you more information about doing this.
We know there is a profile out there. Everybody’s got a profile, and each disease has a profile. We know that. But it’s very easy to make mistakes in establishing what that profile is.
And I’ll give you an example. A little while ago, someone – a colleague of mine – was working to look at diagnosing cancer through breath analysis, went up to the ward and got breath samples from a number of different patients on the ward with this particular cancer; then went around the hospital and found a number of people who didn’t have the cancer and took breath samples from them, looked and thought, ‘Wow, there’s one particular compound in here which is really different in the cancer patients and the controls.’
And they were just about to send this off for publication when someone said, ‘Have you looked at the backgrounds in the ward and the rest of the hospital?’
And it turns out that the actual marker that they thought they’d found for cancer turned out to be something that was present in the air of the ward and not elsewhere in the hospital. So, it’s about understanding all the factors that can actually have an impact on your data.
RICHARD HOLLIMAN
Cool – thanks very much.
CLAIRE TURNER
Thank you.