Skip to navigation Skip to main content Skip to footer

Approved research

Refining polygenic score performance in diverse populations

Principal Investigator: Dr Laramie Duncan
Approved Research ID: 32770
Approval date: July 27th 2018

Lay summary

It has recently been discovered that hundreds (or even thousands) of genetic factors influence traits and diseases like height, blood pressure, and depression. Based on this information, ?genetic scores? can be constructed for each individual. These scores are far from perfect, but they provide some information about a person?s likely traits and risk of certain diseases. However, most research has been conducted in only European ancestry individuals. The purpose of this application is to learn how well these scores work in diverse populations, and to develop methods that make these scores useful for more people. The UK Biobank was designed to ?improve the prevention, diagnosis, and treatment of a wide range of serious and life-threatening illnesses?. In order to ensure that the benefits of research are shared by all, studies must be conducted on the generalizability of results across diverse groups of people. The aims of this project are (1) to identify any ways in which polygenic score predictions may be biased in particular populations, and (2) to develop a list of recommendations regarding implementation and interpretation of polygenic scores across ancestrally diverse populations. Many common traits and diseases (for example: height, depression, and blood pressure) are influenced by hundreds (or even thousands) of genetic factors. ?Genetic scores? can be calculated with this information, and these scores are partially predictive of traits like height and blood pressure. However, these scores work better in some individuals than others. The purpose of this study is to identify ? and fix ? problems with genetic scores, by testing different ways of making genetic scores, and comparing the performance of different scores. We will take a broad approach, and will examine multiple conditions and traits. Full cohort