Genetic testing to check an individual’s risk of developing a variety of diseases has been around for several years. However, up until recently, the information given by this type of testing was unreliable. Testing results from different test providers sometimes gave quite different results and risks. In the US the Federal Food and Drug Administration required 23andMe, the largest provider of this type of testing to stop offering health-related testing until they submitted evidence to the FDA showing the testing was accurate and unlikely to cause harm. The company has done this and since 2015 23andMe and other companies have been offering an increasing menu of health risk and genetic disease carrier status testing.
Increased accuracy in this type of testing requires access to a very large database of genetic sequence information and health records for those who have been tested. The UK Biobank how contains DNA sequence and health information for over 400,000 people. Researchers from the Broad Institute of Massachusetts Institute of Technology and Harvard University have now used this database to construct new algorithms to determine risks of some common diseases.
The researchers developed their algorithms called genome-wide polygenic scores (GPSs) using a subset of 120,280 people from the UK Biobank. They then assessed the performance of their algorithms using an independent testing set comprising the 288,978 participants in the UK Biobank phase 2 genotype data release.
The results of this study showed that the GPSs identified 8% of the population as being at greater than threefold increased risk for coronary artery (heart) disease, 6.1% for atrial fibrillation, 3.5% for type 2 diabetes, 3.2% for inflammatory bowel disease and 1.5% for breast cancer. Current genetic testing looking for rare single-gene mutations that cause some of these diseases such as familial hypercholesterolaemia causing increased risk of heart disease and BRCA mutations increasing risk of breast cancer only identify a tiny proportion of the population as being at this level of risk. The coronary artery disease GPS used information about over six million genetic variants, each of which has only a tiny effect on risk but together identify 20 times more people as being at the same risk as the tiny proportion of people with a mutation in genes causing familial hypercholesterolaemia.
The researchers point out that the GPS could be assessed from the time of birth, well before development of the currently-used risk factors typically used in clinical practice to predict CAD. This could allow individuals identified as being as at increased risk to follow strategies such as adherence to a healthy lifestyle or cholesterol and blood pressure-lowering therapies.
Commenting on the study for theheart.org | Medscape Cardiology, John Mandrola, MD, clinical electrophysiologist, Baptist Medical Associates, Louisville, Kentucky, who was not involved with the study, said he is "optimistic" about the approach. Although it is "still early and there aren't clinical outcomes yet — we need to do these studies — these findings are still important," he said.
While we wait for many years for these clinical outcome studies to become available, developments using this type of technology are not going to stop. It is already known from identical twin studies that many other non-health related traits have a strong genetic component. While intelligence seems to be only about 50% inherited (20% if measured in young children but higher than 50% if measured in young adults), other traits such as height, educational attainment and even income earning potential are quite strongly inherited. If databases like the UK Biobank, but containing this type of information become available it is going to be hard to stop commercial companies offering this type of testing too. Prenatal genetic testing for a limited range of conditions like Down Syndrome is already widely available through NIPT. It is time to start thinking about how far we want this to go.