Modeling Linkage Disequilibrium Increases Accuracy of Polygenic Risk Scores

By Bjarni J. Vilhjálmsson, Jian Yang, Hilary K Finucane, Alexander Gusev, S. Lindström, S. Ripke, G. Genovese, Po-Ru Loh, Gaurav Bhatia, R. Do and 22 more. 2015

1,263 citations recorded by the source corpus.

Subfields: Genetics, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Statistics and Probability

Citations in the corpus

Gene discovery and polygenic prediction from a 1.1-million-person GWAS of educational attainment (2018, methodology): “In practice, this matrix is not known, so we follow Vilhjalmsson et al.(334) and replace the training-sample variance-covariance matrix by an approximation that is estimated using observed LD patterns in a reference sample of conventionally unrelated individuals with European ancestry.”
A tutorial on conducting genome‐wide association studies: Quality control and statistical analysis (2018, positive): “Other programs for the application of PRS are, for example, PLINK ( ‐‐ score ) and LDpred (Purcell et al., 2007; Vilhjalmsson et al., 2015).”
LDpred2: better, faster, stronger (2020, methodology): “The recommended value for this window (in number of variants) has been to use the total number of variants divided by 3000, which corresponds to a window radius of around 2 Mb (Vilhjálmsson et al. 2015).”
Heterogeneity in polygenic scores for common human traits (2017, methodology): “In addition, we were unable to fully evaluate all packages available for creating PGSs because of the lack of detailed GWAS meta-analysis results available from consortia that are required in these packages, particularly LDPred (7).”
Multi-trait analysis of genome-wide association summary statistics using MTAG (2018, positive): “We calculated the SNP weights using the software package LDpred, assuming a fraction of causal SNPs equal to 1.”
The critical needs and challenges for genetic architecture studies in Africa (2018, positive): “performing the worst in African Americans (and by extension, likely even worse in 18 eastern, central, and southern African populations) [83-86].”
A machine-learning heuristic to improve gene score prediction of polygenic traits (2017, methodology): “β β = − ˆ ˆ d w ( )sign ( ) (4) ext ext”
PRSice-2: Polygenic Risk Score software for biobank-scale data (2019, neutral): “However, the calculation of PRS, which involves parameter optimization [13–16], can be a computationally intensive process, especially for large datasets and when multiple analyses are conducted.”