An key approach is to use polygenic risk scores (extensively reviewed and discussed elsewhere32). A polygenic risk score (PRS)33 is an approximate measure of an individual’s common variant genetic propensity for a given disorder and, at a population level shows some predictive power34 for case-control status. PRS approaches provide several potential routes to drug development, including identification of genetically associated endophenotypes and biomarkers. PRS can also be exploited to improve clinical trial efficacy. Super controls can be chosen by selecting participants with very low PRS for the disease, or PRS for low risk of side-effects or where differential diagnosis is unclear. This may convey particular benefit in trials for diseases such as Alzheimer’s (being investigated by a new workgroup in the PGC), where defining cases and controls is challenging. Furthermore, prevention trials could enlist high risk individuals from the top end of the PRS distribution35, which, amongst other benefits, may be less expensive and confounded than the sibling design36. Current studies in psychiatry are attempting to improve prediction of diagnosis or treatment response, for example in first episode psychosis37.
