Skip to navigation Skip to main content Skip to footer

Approved Research

Composite polygenic risk score analysis of trans-ethnic portability and within-family validation of prediction, using compressed sensing and high-dimensional statistical methods in complex traits.

Principal Investigator: Mr Laurent Tellier
Approved Research ID: 54679
Approval date: October 19th 2022

Lay summary

Our goal is to develop and test new computational methods for determining the genetic contribution to complex traits, including highly heritable conditions such as Diabetes, Alzheimer's Disease, Breast Cancer, Schizophrenia, and Major Depressive Disorder. Most traits are caused by both environmental components and a genetic component. The latter is usually the sum total of the contribution from many (up to tens of thousands of) genes.Through machine learning algorithms, we use DNA-information alone to predict traits. These disease risk predictions are not perfect, but can be of huge importance when integrated into clinical treatment and prevention. For example, genetic breast cancer prediction can inform women at extra high risk to begin early screening, and can thus alone save thousands of lives yearly.

Our research is both focused on producing the best possible predictors for the most important diseases, as well as for people of all ancestries, via further methodological development. We are also advancing the research on how to simultaneously combine multiple disease risk predictions into genomic indices, and how to best deploy such indices in a clinical application context.

There is already a real impact through the early clinical adaptations, and the coming health benefits for both individuals and the society at large are enormous. This project has a commitment to advance the field as long as there are scientific gains to be made for both a deeper understanding of the genetic nature and the best possible predictors of complex traits.