Announcing a meeting of the Statistical Society of Australia, W.A. Branch.
6:00 ᴘᴍ on Tuesday 12th November 2019
BIOLOGICAL SCIENCES LECTURE THEATRE (240.2.051)
Using polygenic risk scores to predict risk of disease: hype, hope and statistical reality
Professor of Genetic Epidemiology & Statistics
King’s College London
Most common disorders such as depression, breast cancer and coronary heart disease have a genetic component, which has long been observed from increased rates of disease in family members. Genetic studies have shown a polygenic component to these diseases, which is due to the cumulative effect of hundreds and thousands of genetic variants, each having a minor effect on disease risk. Individual-level polygenic risk scores can be constructed as a measure of genetic liability of disease, as a sum of the genetic loading, multiplied by the risk conferred by each variant. By the central limit theorem, polygenic risk scores have a normal distribution, and are often standardised to the population mean. However, polygenic risk scores differ by ethnicity, and principal components of genetic ancestry are necessary as covariates to control for this.
Polygenic risk scores may be applied in two different settings:
- In research studies, to test whether disease cases have higher scores than controls, and as a tool to characterise heterogeneity within cases. Logistic regression is used as an analysis tool, with disease status (affected, unaffected) as outcome variable, polygenic risk scores as explanatory variable, and ancestry-informative principal components as covariates. Summary measures for this include the improved model fit, from change in Nagelkerke’s R2, and the AUC (Area under the curve) for predictive ability.
- The ultimate utility of polygenic risk scores would be to use individual polygenic risk scores, to identify those at high risk of disease, enabling them to enter relevant screening or intervention programmes.
In this seminar, I will discuss the current utility of using polygenic risk scores to predict disease risk (currently very low), and the potential for increasing their predictive ability.
ABOUT THE SPEAKER:
Cathryn Lewis is Professor of Genetic Epidemiology & Statistics at King’s College London, where she leads the Statistical Genetics Unit. Her academic training is in mathematics and statistics, and she has been involved in genetic studies since her PhD. She co-chairs the Psychiatric Genomics Consortium Major Depressive Disorder Working group, and leads the NIHR Maudsley BRC Biomarkers and Genomics theme. Her multi-disciplinary research group identifies and characterises genetic variants conferring risk of disease, including depression, schizophrenia, and stroke. A major focus is risk assessment, determining how the polygenic component of mental health disorders can be measured accurately and communicated effectively. Cathryn has a B.A. from St Hilda’s College, University of Oxford and a M.Sc. with distinction and a Ph.D. from The University of Sheffield.
Cathryn Lewis is the SSA WA Branch Hansford-Miller Fellow for 2019.
Members and guests are invited to mingle over wine and cheese from 5:30 ᴘᴍ onwards in the Biological Sciences Lecture Theatre. Following the meeting you are invited to dine with the speaker at Bateman Chinese Eating House & Take Away; please RSVP to Brenton Clarke B.Clarke@murdoch.edu.au by 11 am Tuesday 12th November. Visitors are welcome.
Parking is free on the Murdoch University campus after 5:00 ᴘᴍ. We suggest parking in Car Park 2 (searchable in Google Maps) that can be entered from campus entrance C on South St.
For further information please contact the Branch Secretary, Rick Tankard, Murdoch University.
He can be reached by email at email@example.com or by phone at (08) 9360 2820.