F062 Clustering of disease progression trajectories in Huntington’s disease psychosis and non-psychosis: an investigation into the predicting features and genetic modifiers

<h3>Background</h3> Defined as the presence of hallucinations and/or delusions, psychosis in Huntington’s Disease (HD) has an estimated prevalence of 3–11%. HD-psychosis has been repeatedly linked to lower functional ability, greater motor disturbances, longer illness duration and reduced cognition. While there is evidence of a unique clinical course, it remains unclear whether these patients experience a unique progression trajectory (i.e. rapid, moderate, slow) compared to non-psychosis cases. <h3>Aims</h3> This research intends to model and compare the longitudinal disease progression trajectories of HD patients with psychosis and without psychosis. We aim to explore unique characteristics, and genetic modifiers that may drive these distinct trajectories (if any) between HD-psychosis and HD non-psychosis. <h3>Methods</h3> Taking a novel approach, our research project will implement k-means clustering to HD-psychosis and HD non-psychosis cases both collectively and independently to identify groups of slow, moderate, and rapid progressors. To identify the influencing features of the groupings, a supervised machine learning technique like XG Boost will be applied. Known genetic modifiers identified by previous HD genome wide association studies (GWASs) and a schizophrenia polygenic risk score (PRS) based on previous schizophrenia GWAS will be employed in our investigation of genetic modifiers. A stratified analysis will be used to investigate the moderating effects of the identified genetic loci and PRS.