Approved Research

Polygenic prediction and interaction modeling for blood pressure traits and coronary artery disease using diverse machine-learning methods.

Lay summary

Scientific rationale: High blood pressure is the leading preventable risk factor for cardiovascular disease such as heart attack and all-cause mortality worldwide. It is known that it is partly genetically determined, and hundreds of genetic variants are known to play a role to determine blood pressure. In addition, it is assumed that variants also depend on each other in their effects, which is not mirrored in currently available genetic scores.

Aims: The aims of the proposed project therefore are to firstly develop genetic scores for blood pressure that make use of genome-wide data but also take interactions between genetic variants into account. To maximally use the available data of the UK biobank, different machine-learning methods will be used. Secondly, a specific novel machine-learning method will be applied to identify interactive effects specifically on CAD to validate preliminary findings from previous analyses in a large German case-control data set.

Project duration: 36 months

Public health impact: Genetic prediction of blood pressure may be improved and the genetic background of CAD may be understood on a more detailed level than before, thus guiding further studies on the prevention of cardiovascular disease.