Can you tell from your DNA which diseases you are at greater risk of, and how is that measured?
Your DNA and certain proteins in your blood provide a useful picture of your predisposition to heart disease and other common conditions, but in people of non-European ancestry those scores are less reliable.
Yes, you can. The most widely used method is the polygenic risk score: the effects of sometimes more than one million small DNA variations are added together to produce a single number. The higher that number, the greater your genetic predisposition. For heart disease, the 20% with the highest score had a more than fourfold higher risk than the 20% with the lowest score1. For five conditions at once, including atrial fibrillation, type 2 diabetes and breast cancer, such a score could identify 1 to 8 percent of the population with a more than threefold elevated risk2.
There are, however, two serious limitations. Most scores have been built using data from people of European ancestry. In people with a different background they are less accurate and can sometimes even be misleading3. In addition, they always express probabilities, never certainties: a high score does not mean you will definitely develop the disease.
Besides DNA, you can also measure proteins in your blood. Proteins are the substances your genes produce; their presence reflects what is happening in your organs. Measuring 5 to 20 proteins turned out to predict the 10-year risk of 67 different diseases better than standard medical information alone4. Using the same approach you can also see which organs are ageing faster than expected. Nearly 20% of people have one organ that is ageing at a strongly accelerated rate, which increases the risk of death by 20 to 50%. In the case of accelerated heart ageing, the risk of heart failure rose by as much as 250%5.
DNA and proteins together produce even sharper predictions. In people with type 2 diabetes, a model based on nine proteins linked to both diabetes and heart disease predicted future heart disease better than DNA or medical data separately6.
A completely different application is reading the full DNA of newborns to detect rare, serious mutations in a single gene. This can be useful, but interpretation is complex: by no means do all variants found have a clear meaning, and it is ethically difficult to decide what to report back to parents when risks only become relevant later in life7. Research on this point is still at an early stage.
Based on 7 claims with PMIDs: 30309464, 30104762, 31765077, 39039249, 38057571, 40087642, 30576641. Three claims rest on large cohort studies (strong quality of evidence), two on moderate evidence, two on moderate evidence with ethical considerations. All findings are associative; none of the DNA or protein measurements have been proven causal in an interventional sense.