Rxivist logo

Positioning Personal Polygenic Risk score against the population background.

By Ganna Leonenko, Emily Baker, Karl Michael Schmidt, Valentina Escott-Price, the Alzheimer’s Disease Neuroimaging Initiative

Posted 24 Oct 2019
bioRxiv DOI: 10.1101/813170

The polygenic risk scores (PRS) approach has been widely used across different traits for estimating polygenic risk, pleiotropy and disease prediction, but mostly in European populations. The predictive ability of the PRS in non-European populations is currently limited due to the lack of genetic research performed in populations of non-European ancestry. One of the main challenges of the practical use of PRS is to place an individuals personal score in the context of the PRS distribution in the underlying population. In this paper we present an approach for estimating the parameters of the PRS distribution in a population using summary information from public data. Unstandardized PRS are usually not directly comparable even between European studies. Our approach can be used for standardisation whilst accounting for genotyping platforms, data quality and ancestry. It can be applied to assessing polygenic disease risk for individuals from a European population for any complex genetic disorder and, assuming that most of the disease risk loci are likely to be shared between populations, to estimating the disease risk for individuals from other populations. We demonstrate the precision of our method with simulations. We show the utility of our estimates in application to Alzheimers disease in the Alzheimers Disease Neuroimaging Initiative (ADNI) study. We present population specific PRSs for different populations using 1000 Genomes data.

Download data

  • Downloaded 407 times
  • Download rankings, all-time:
    • Site-wide: 68,254
    • In bioinformatics: 6,554
  • Year to date:
    • Site-wide: 81,530
  • Since beginning of last month:
    • Site-wide: 116,613

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide