Contribution of GWAS hits, low penetrance genes, high penetrance genes, and high penetrance variants to prediction results.

<p>Each row represents a phenotype predicted with AUC>0.7 by genome sequence (<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1003825#pcbi-1003825-g002" target="_blank">Figure 2</a> (b)) and contains five cells. Cells are colored by the area under the ROC curve (AUC) yielded by a model that contains only 1:GWAS hits, 2:Low penetrance genes, 3:High penetrance genes, 4:High penetrance variants. The fifth cell shows AUC of the combination model used to assess results in this work that considers all of 1,2,3, and 4. The combination model generally yields the best performance: however, for most phenotypes, only one or two of 1, 2, 3 or 4 appears to contribute.