Polygenic risk scores and DTC testing: a problematic pairing?
We explore polygenic risk scores and why their application in at-home genomic testing has been brought into question
A large British Medical Journal study has questioned the usefulness of polygenic scores for predicting disease risk, with results suggesting an effect on healthcare “disproportionate to their performance”.
In this blog, we explain the findings and their practical implications, especially when used in direct-to-consumer testing, and introduce an NHS-partnered research programme that aims to develop a new, more reliable type of risk score.
What are polygenic risk scores?
Historically, genetic medicine has focused on heritable, often rare, conditions caused by variants in single genes. These variants result in a person developing a particular condition, or having a very high chance of doing so.
Genome-wide association studies (GWAS) assess a large number of variants throughout the genome for an association with a particular trait or condition, usually common conditions that are known to have many genetic and environmental factors influencing their development such as diabetes, cancer and cardiovascular disease. Variants that are identified in this way often only have a very small effect on an individual’s chance of developing the condition and can only be identified in studies using data from thousands of individuals.
By bringing together the combined effects of these variants, polygenic risk scores aim to quantify the total impact of these gene variants in a person’s genome, and use this information to understand their overall chance of developing a particular condition, compared to the population at large. This is a comparatively new area of genomic science, and one that is still being evaluated to understand whether, and how, this type of genomic data can be harnessed to improve people’s health.
Direct-to-consumer tests may be misleading
The BMJ paper detailed research carried out at University College London. They looked at a catalogue containing more than 900 different polygenic risk scores for 310 different diseases.
Lead author Professor Aroon Hingorani from UCL’s Faculty of Population Health Sciences said: “We found that, when held to the same standards as employed for other tests in medicine, polygenic risk scores performed poorly for prediction and screening across a range of common diseases.”
Examples of this are polygenic scores for breast cancer and cardiovascular disease. The study found that, on average, they only predicted around 10% of breast cancer cases, and 12% of people who would go on to develop coronary artery disease. They also predicted illness for 5% of people who would not ultimately be affected.
These scores form the basis of many commercially available tests that make bold claims about their predictive power.
The authors of the BMJ research suggest that “policy makers might wish to consider stricter regulation of commercial genetic tests based on polygenic risk scores, with a focus on clinical performance … to protect the public from unrealistic expectations and already stretched public health systems from becoming overburdened by the management of false positive results.”
Our Future Health could tell us more
Our Future Health is a research programme being run in partnership with the NHS. It will recruit 5 million volunteers who will share their data to help scientists understand, treat, and prevent common conditions such as cancer, dementia, diabetes and cardiovascular disease.
One of the project’s aims is to evaluate whether risk scores informed by genomic data can help predict and prevent common diseases, enabling people to live healthier lives for longer.
The project’s chief medical officer, Dr Raghib Ali, said: “This research study rightly highlights that for many health conditions genetic risk scores alone may have limited usefulness, because other factors such as deprivation, lifestyles and environment are also important.”
He explained that Our Future Health plans to develop ‘integrated risk scores’ that include genomic and non-genomic risk factors:
“We hope these integrated risk scores can identify people more likely to develop diseases, but this is a relatively new area of science and there are still unanswered questions around it.”
