I can claim a degree of novelty – my test was carried out by the Icelandic genomics giant deCODEme, rather than the more ubiquitous 23andMe. I also promise to keep nothing back, I have copied every shred of my raw data directly into this post.
Given my DNA is in part owned by my family, I was not planning to divulge all the results. But investigating the raw data has made me comfortable that what lies underneath holds no great secrets.
I took the cancer scan, which generates risk assessments for seven common cancers. The results are revealed one at a time, like flipping cards in a hand of morbid poker.
Here is the first one:
The 1.0 in the middle of the bar is average risk, anything lower and greener is good, to the right and red is bad. So this is good, and as I click through the list my heart rate increases and my palms become clammy.
I feel like I’m on a roll now, how long will the luck last?
A great result as I have a family history of breast cancer.
I’m now aiming for a full house, so the next one is a letdown:
I hadn’t picked up that being female I would only get six and not seven results. I had even spent time thinking through how a high risk prostate cancer value could affect my close male relatives.
The rush of luck is over - the remaining three results sit close to the centre line (although I struggle to gain any extra insight from the numbers being quoted to two decimal places):
Each result was preceded by a critical health information page which I dismissed without reading (the first time guiltily, subsequently with abandon). My brain was viscerally primed to uncover my cancer risks and I was too impatient to digest details.
Later, when I had calmed down, I had to hunt around to trace the information again, and the option of accessing it straight after the results would have been useful. This illustrates a key dilemma in dispensing genetic information direct to consumers – how do you ensure they read the small print?
I was keen to download my raw data to see what was behind the headlines. After spotting a warning that scan results could contain one million lines of data, I pressed the download button and went to make a cup of tea. The file downloaded instantly, here it all is:
In terms of base pairs for your buck, the test does not score highly, and it won’t be long before the $500 USD price tag will buy me all my 3 billion bases rather than just 33 select ones.
The test's real value comes from areas in which deCODEme excel: the painstaking discovery of meaning hidden in sequence muddle and the careful communication of the resulting risk factors.
Finding even a handful of SNPs that reproducibly effect cancer risk is a tour de force of genetics. When I drill down to the peer reviewed papers behind each data point I can see that deCODE genetics have carried out the bulk of this work. The scientific details reveal a mountain of research, with data from tens of thousands of people used to define each SNP.
To understand the significance of my lower risk results I need to take the population frequency of each cancer into account. For both bladder and breast cancer my lovely green arrows translate to only a small decrease in absolute risk. Combined with the three results clustered around 1.0, this makes me a resounding Mrs Average for five out of six of the cancer types.
Only basal cell carcinoma is left. This gave my biggest risk swing and is the most prevalent of the cancers tested, affecting 25% of the population. The results reduce my personal risk to 16.5% – the only number in the whole test that stands out.
The decreased risk is reassuring but will not make me lax about using sunscreen. If anything the numbers have made me more vigilant – I didn’t realise that basal cell carcinoma was so prevalent. And despite this being is my lowest relative risk, the high background level means it is still my highest absolute risk.
However, even for the basal cell result, any sense of deviation from the mean disappears as I drill down to the three constituent SNPs. All are pretty average too – each present in more than 40% of the population. Individually they confer a slight decrease in risk, and it is only their combination by multiplication that gives an overall risk well below average.
By exploring my data I feel like I’ve descended from the heightened emotion of the headline results to a baseline of bog standard DNA. It is the very ordinary nature of the data that makes me happy to share everything here.
Although it is boring to be average, this is one time I’m thankful to be part of the crowd. Middling results are by no means guaranteed – you could learn you have 2.3 times the background risk of breast cancer (giving a lifetime risk of 28%) or over four times that of prostate cancer (lifetime risk of 66%). For a minority of customers, the results could be far from mundane.
For me, the process of taking the test has been much more educational than the test results themselves. The deCODEme website is full of well presented pre- and post-test information, and their email support has been prompt and factual. The scan has made me mull over ownership of my DNA and I’ve been inspired to look into my family history.
Even if genomic fatigue sets in and we tire of hearing about other people’s genotyping results I don’t believe we will ever stop being fascinated by our own, even the bits that are rather bland.
A review of the Association of British Science Writers genetics testing conference session
Questions to ask before buying an internet genetics test
The test was donated by deCODE to the Association of British Science Writers for their annual conference.