Genome sequences and electronic health records have been combined to help identify those at risk of a life-threatening rupture of the aorta.
Genetic factors are known to contribute to the risk of abdominal aortic aneurysm, a fragile bulge in the lower section of the aorta, the body’s main artery. But the condition is complex and difficult to study.
Philip Tsao and Michael Snyder at Stanford University in California and their colleagues sequenced the full genomes of 268 people with the condition, and 133 controls. They then used a machine-learning method to analyse the genetic data, and built a machine-learning model based on information from electronic health records, such as cholesterol levels. Finally, the team designed a model based on both genetic data and health records that made highly accurate predictions of which samples came from people with the disease.
The genetic analysis singled out 60 genes that were more likely to carry mutations in individuals with abdominal aortic aneurysms than in the controls. These genes tended to be expressed at higher levels in tissue taken from people with the disease, compared to those without it.
Correction: A previous version of this story attributed the work to only the corresponding authors.