Using artificial intelligence to predict the underlying causes of high blood pressure

The BHF-Turing Cardiovascular Data Science Awards (Second Call): Molecular causal networks of hypertension – a machine learning approach (joint funding with The Alan Turing Institute)

High blood pressure (hypertension) is a major cause of heart attacks and strokes. Studies of the human genome have identified over a thousand differences between people’s DNA (genetic variants) that are associated with blood pressure. The way these variants influence blood pressure and the risk of hypertension are poorly understood. Does simply switching genes on or off cause high blood pressure? Or do they trigger a chain of molecular events, for example changing the amount that other genes are switched on, and therefore causing high blood pressure indirectly? Statisticians and physicians at the University of Manchester will attempt to answer these questions using a type of artificial intelligence called machine learning. Machine learning is well-suited to complex biological systems because it can efficiently identify the network of relationships between many factors. The team will use machine learning to explore the relationships between genetic variants associated with blood pressure and the risk of hypertension. A large volume of data from 500 human kidneys (an organ which plays a central role in regulating blood pressure) and existing data from large-scale genetic studies will be used to develop the method. Identifying the network of factors that cause hypertension will inform the development of new treatments. This new artificial intelligence could also be used to understand the causes of other diseases.