Fig. 4
From: Accounting for overlapping annotations in genomic prediction models of complex traits
Impact of BayesRC random annotation assignment on large-effect QTL variance estimation. Boxplots represent the difference with BayesRC in estimated posterior variances for large-effect QTLs for BayesRC\(\pi\) and BayesRC+. Results are shown according to the randomly assigned annotation categories (strongly, moderately, weakly/unenriched) used for BayesRC. All large QTLs across the 50 independent datasets are represented (\(h^2=0.5\), \(k_\text {large}=1\%\), scenario D with 9 annotations)