Fig. 2
From: Accounting for overlapping annotations in genomic prediction models of complex traits
Using the PAIP to interpret annotation importance for BayesRC\(\pi\). The PAIP, “Posterior Annotation Inclusion Probability”, is defined as the frequency of marker assignment in each annotation across iterations. Results are shown for \(h^2=0.5\), \(k_\text {large}=1\%\), and scenario A. A Posterior mean frequency of marker assignment to the strongly enriched annotation (i.e., strongly enriched PAIP) by simulated effect size category (null, small, medium, high). Results are averaged across 50 independent datasets. B Distribution of the log posterior variance of markers by PAIP-assigned annotation for one illustrative dataset