Figure 2 presents the Q-Q plot for P3 amplitude, while Figure 3 presents the Q-Q plot for the genetic factor score. Substantial deviation from the expected line representing the null distribution can indicate inflated (or deflated) power, such as might result from population stratification. This was not the case, an inference corroborated by the genomic control statistics, which were close to 1 for both measures (with 1 indicating that the observed values conform exactly to expectation under the null hypothesis). Our analytic approach therefore appears to have appropriately accommodated the lack of independence in our family data, and there was no meaningful residual ethnic stratification. There was also no evidence for significant associations. None of the SNP effects on either P3 or the genetic factor score was genome-wide significant (p values > 5 × 10−8). An apparent excess of small p values, which appear as large values of −log10(p), was evident in the Q-Q plots, especially for P3 amplitude. These subthreshold p values may indicate that there are true associations hidden in the GWAS signal. We list the results for
