In recent years, scientists have been working on developing polygenic scores, which are calculations that determine a person’s likelihood of developing a particular disease based on numerous genetic variations throughout their genome. While the accuracy of these scores has improved for certain diseases and specific populations, they have been less reliable for individuals of non-European ancestry. This is primarily due to the fact that the genetic datasets used to create these scores have primarily consisted of data from individuals of European descent.
However, a team of researchers from the Cardiovascular Disease Initiative at the Broad Institute of MIT and Harvard, along with Massachusetts General Hospital (MGH), has introduced a new approach that significantly enhances the accuracy of genetic risk prediction for heart disease across all ethnic backgrounds.
To develop their polygenic score, the scientists utilized data from genetic studies involving over 1 million individuals. In addition, they incorporated genetic variations associated with 10 related traits, including blood pressure and body mass index, into their calculations to further refine the score. The newly developed polygenic score surpassed all existing scores in predicting the risk of coronary artery disease, which is the leading cause of death globally, for participants of African, European, Hispanic, and South Asian ancestry.
This innovative approach holds the potential to enable clinicians to identify individuals at high risk for heart disease earlier in life, possibly even from birth. This early identification could facilitate the implementation of interventions such as cholesterol-lowering medications or lifestyle modifications, which have been proven to mitigate and normalize the heightened genetic risk. Published in Nature Medicine, the study’s findings suggest that this framework could be extended to enhance genetic risk prediction for other diseases and traits as well.
Amit V. Khera, co-senior author of the study and formerly associated with the Broad as a Merkin Institute Fellow and currently serving as the vice president of genomic medicine at Verve Therapeutics and a cardiologist at Brigham and Women’s Hospital, emphasized the significance of being able to identify genetic risk early in life, even from birth. By employing larger and more diverse datasets, the newly developed polygenic score can more effectively pinpoint individuals at high risk who might otherwise go undetected. Pradeep Natarajan, co-senior author and associate member of the Broad, director of preventive cardiology, the Paul & Phyllis Fireman Endowed Chair in Vascular Medicine at MGH, and associate professor of medicine at Harvard Medical School, added that their score has the potential to identify high-risk individuals more accurately by leveraging a more comprehensive and diverse dataset.
Prediction progress
To develop the new polygenic score, the researchers collected data from over one million individuals, including nearly 270,000 individuals diagnosed with coronary artery disease. This dataset was significantly larger and more diverse compared to their previous study in 2018, which only included tens of thousands of individuals with the disease. The team was able to incorporate data from studies with individuals of African, Hispanic, and South Asian ancestry, such as the U.S. Veterans Affairs Million Veteran Program, which provided a more comprehensive representation of genetic diversity.
The 2018 scores, which were primarily based on European populations, did not perform well in predicting risk for individuals of other ancestries. Recognizing this limitation, the scientific community has been focused on improving risk prediction across different ethnic backgrounds.
In order to capture the influence of genetic variants with smaller impacts on heart disease risk more accurately, Minxian (Wallace) Wang, a co-first author of the study and former computational biologist at the Broad, developed a pipeline. This pipeline prioritized DNA changes that are known to affect both heart disease risk and related traits like body mass index, smoking status, and blood pressure. Notably, approximately half of the predictive power of the new score came from heart disease studies themselves, while the other half came from studies on these related risk factors.
When applied to a separate dataset comprising individuals from diverse ancestries, the newly developed score, named GPSMult, identified a greater number of individuals at both the highest and lowest risk of heart disease compared to previous scores. For instance, individuals in the lowest percentile of the score had less than a 1% chance of being diagnosed with heart disease by middle age, whereas those in the highest percentile had a 16% chance. Interestingly, the team was able to identify 3% of unaffected individuals who, based solely on common DNA variations, had a risk for future cardiac events, such as heart attacks, equivalent to individuals who had already been diagnosed with the disease.
These findings highlight the advantages of incorporating multiple traits and utilizing multi-ancestry data when calculating polygenic risk scores. Furthermore, the approach holds promise for improving risk prediction for other diseases. The researchers plan to refine their method by incorporating even larger and more diverse datasets, employing new computational approaches that account for complex genome architecture, and integrating clinical risk factors to enhance the relevance of the scores for physicians.
Amit V. Khera emphasized that there is still room for improvement in genetic predictors for heart disease, and future tests will likely become even more effective. Additionally, efforts are needed to determine the best way to integrate these tests into clinical practice and make them a standard part of healthcare.