Skip to main content
. 2017 Sep 9;216(9):1063–1069. doi: 10.1093/infdis/jix470

Figure 2.

Figure 2.

Association results of gene-based testing. Quantile–quantile plots of observed −log10(P) (diamonds, y axis) vs −log10(P) expected under the null hypothesis (red dashed line). Analysis was performed including all variants regardless of predicted function (A), only variants predicted to modify the protein (B), and only variants predicted to highly impact the protein structure (C). Only genes within the MHC were associated after correcting for multiple comparisons. The distribution of P values of non-MHC genes (blue diamonds) closely follows the expected null distribution.