There is great promise in personalised medicine. For example, given an individual's genome, it should in theory be possible to accurately predict their risk of developing diabetes; or if the individual is diagnosed with epilepsy, we should be able to use their genetic information to determine best course of treatment. But despite the ever-increasing amounts of genetic data being produced, we currently lack statistical tools to efficiently analyse these data and make personalised medicine a reality. In this fellowship I will develop novel statistical methods for analysing genetic data, then apply these methods to large-scale datasets. I will tackle important problems in medical genetics, such as constructing models to predict disease susceptibility and severity, and classifying heterogeneous neurological diseases. Finally, I will make all my methods freely-available, so other researchers can apply them to their datasets.