High street genetic tests – Dr Michael Mosley reports
You can buy a home genetics test kit from companies like 23andMe for just £125, but can they really offer us a useful peek at our future health problems, or is DNA testing a waste of time and money?
The first genetics test I ever did was while I was making a programme for Horizon called The Truth About Exercise back in 2012. The premise was that I would test the claim that you can get most of the benefits of exercise from just three minutes a week of high intensity training (HIT).
My guide into the world of HIT was Prof Jamie Timmons, who then worked at Birmingham University as an expert in ageing biology. Despite my scepticism, Timmons assured me that if I did just a few minutes of HIT a week for four weeks, I could expect to see significant changes in a couple of important health indices.
The first improvement I could expect to see was in my insulin sensitivity. Insulin removes sugar from the blood and it controls fat. When it becomes ineffective, you become diabetic. The second improvement I was likely to see would be in my aerobic fitness (VO2 max). Aerobic fitness is a measure of how good your heart and lungs are at getting oxygen into your body, and is an excellent predictor of future health. It is usually measured in an exercise laboratory, though there are other ways of assessing it.
Before I did his HIT programme, Timmons gave me a couple of test tubes to spit into. I knew it was for some form of genetic test, but the BBC producer asked Timmons not to tell me any more than that. So I went off and dutifully did my four weeks of HIT, and then went back to the lab to be tested again.
The results were mixed. My insulin sensitivity improved by a remarkable 24 per cent, which was extremely satisfying, but my aerobic fitness did not improve at all. Timmons was not surprised. It turns out that the genetic test I had done suggested that I would be what’s called a ‘non-responder’ when it comes to aerobic fitness. However much exercise I had done, and in whatever form, my VO2 max would not have improved. My dreams of winning Olympic glory ended there and then.
In your genes
It’s been known for some time that people respond to exercise regimes in very different ways. Scientists had previously assumed that this is because of differing levels of compliance: high compliers get fitter and poor compliers don’t. However, it is now clear that genetics plays a big part.
When Timmons and his collaborators looked at studies they found that 15 per cent of people see big improvements in VO2 max when put through a supervised exercise regime, so-called ‘super-responders’. On the other hand, 20 per cent show no real improvement at all – ‘non-responders’ such as me. They investigated the reasons for these variations and discovered that much of the difference could be traced to a small number of genes. So what’s the point of doing a test like this?
“If you don’t normally do much exercise and you take the test and find out that you are a high-responder, then hopefully you will be motivated to get up and get going,” Timmons said. “If you are a non-responder, then you know you need to adapt your workout routine to make it more effective, perhaps have longer periods of recovery.”
One issue with all genetic tests is reliability. How do you know that the answers are valid? I trust Timmons because his research is based on good science, but there are plenty of other genetic tests that are not. A friend of mine, Peta Bee, who writes about exercise for The Times, decided to give it a go. Unlike me, she is very keen on exercise and likes nothing better than a long run. She sent a swab off to Timmons and the answer she got back was the one she was expecting: she was a high aerobic responder, best suited to endurance-type activities.
But when she sent a swab off to another company, they told her that her endurance potential was low and she was much better suited to power and strength activities such as sprinting and weight training. They also recommended netball and football, neither of which has she ever shown the remotest enthusiasm or aptitude for. Timmons is not surprised because, as he put it, “lots of companies are offering genetic tests for sporting performance and personalised macronutrition, but most are completely implausible.”
Wind forward a few years and Horizon once more asked me to donate spit, this time to try out a genetic test provided by 23andMe. The Californian company is named after the 23 pairs of chromosomes in a normal human cell. The US Food and Drug Administration (FDA) decided a few years ago that they and other similar companies were selling ‘medical devices’ and told them to stop marketing the health side of their product. So they stopped advertising and providing health-related information in the US, but continue to offer genetic ancestry details and raw data.
23andMe certainly make genetic testing very simple. I logged on and paid £125. You can also buy the kit at a high street chemist. A short while later they sent me a package with instructions.
I provided a spit sample into a tube included in the kit and sent it back. A few weeks later the results pinged up on my computer. The website is rather impressive. They give you lots of data about your genome, along with references to the studies that form the basis of their claims. They also include a star marking relating to how reliable the data is likely to be.
I started by looking at my ancestry, and discovered that I am 98 per cent European, with just a hint of the Middle East and North Africa – 1 per cent. I am also 1 per cent Asian. This fits in with what I know about my family tree. I had a quick look at inherited conditions and was relieved to see that I am not a carrier of any of the genetic mutations they list, including cystic fibrosis.
I then opened the section on traits. These include things like alcohol flush reaction. Some people, particularly those from South East Asia, flush when they drink alcohol. This is largely down to their reduced ability to metabolise alcohol, which in turn is down to the effectiveness of genes that produce the enzyme alcohol dehydrogenase. I don’t flush. They were very confident I would have straighter hair than average, which is right, but said it was only 28 per cent likely that it would be blonde. I’m not. They told me I’d be tolerant to lactose, which I am, but said my muscle performance suggested I would be a sprinter, which is not true.
I moved on to genetic risk factors, which is undoubtedly the most controversial part of the test. I was particularly interested in my risk of getting Alzheimer’s, as I suspect my father was becoming a little bit demented towards the end of his life. There is one gene in particular, ApoE, that is strongly implicated in late-onset Alzheimer’s (after the age of 65). No-one knows quite how it works, but ApoE influences the build-up of a protein named amyloid beta in the brain. This protein is found in higher levels in people with Alzheimer’s.
The 23andMe test covers three ApoE variants: e2, e3 and e4. It is the e4 variant of the gene that you want to avoid.
According to the website, if you are of European ancestry then one copy of e4 means you have an 18 to 35 per cent of developing Alzheimer’s disease by the age of 85. These numbers rise to 51 to 68 per cent if you have two copies of the e4 variant. So, did I really want to know? After a brief pause I decided the answer was ‘yes’, so I downloaded my results. I was relieved to get the following message: “Michael Mosley has two copies of the ApoE e3 variant, this variant is not associated with a high risk of developing Alzheimer’s disease.”
I don’t know how I would have responded if the answer had been ‘higher risk’. According to the website, having the e4 variant does not mean that a person will definitely develop Alzheimer’s. But I’m not sure how reassuring that would be.
I went to visit geneticist Dr Ewan Birney, Director of the European Bioinformatics Institute, to see what he thinks about these sort of tests.
“I’m not a fan,” he said. “There is a fun side to it, you can use it to trace your ancestry, but I wouldn’t recommend it as a way of monitoring my health or your health. There’s a risk you might get worried inappropriately, that you will become obsessed about a particular diagnosis that may or may not be right. So this is best handled when you’re meeting a clinician who has experience of all of these scenarios to give you good advice.”
Birney said the tests are quite reliable when it comes to conditions that are affected by a single gene mutation, but not for many common disorders.
“For most common diseases, things like heart disease and Type 2 diabetes, your doctor will know a lot more just by doing some simple tests and finding out about your family history,” he said. “How heavy you are, whether you smoke, whether your parents had heart disease – these are more predictive than a spit test. And the advice you’re going to get from your doctor is going to be the same regardless of your genetics, which is don’t eat so much, don’t smoke, exercise more.”
Birney thinks there is a place for personalised genetic testing, but within the healthcare system. He gave me the example of warfarin. Warfarin is an anticoagulant that’s given to patients with an irregular heartbeat or deep vein thrombosis. The problem is that people metabolise warfarin at very different rates, so it is hard to know what dose to put a patient on. Doctors normally start on a low dose, test your clotting time, then increase it gradually. A genetic test that can reliably predict how quickly you will metabolise warfarin would mean you could be started on the right dose quickly, saving lives and a huge amount of testing.
So is this the future, I asked? “No,” he said. “It’s not the future. It’s now.”
Dr Michael Mosley is a BBC presenter and journalist.
This article has been edited for the web. The original version of this article appears in the September 2015 issue of BBC Focus magazine – subscribe here.
(This article was corrected on 22 September 2015 to clarify which services 23andMe are selling in the US )