Fig 3. Influence of filtering threshold on the power to detect GxS.

Shown is the power to detect GxS for the same approaches as in Fig 2 (n₁ = 100,000, n₂ = 100,000), but here with varying filtering thresholds and fixed $R_2^2$ relative to $R_1^2$ for different types of GxS: A. qualitative GxS with small stratum-specific effects ($R_1^2=0.014\%,R_2^2=0.014\%$ into opposite direction), B. pure GxS with medium sized stratum 1 effect ($R_1^2=0.058\%,R_2^2=0\%$), and C. quantitative GxS with large stratum 1 and smaller stratum 2 effect ($R_1^2=0.167\%,R_2^2=0.042\%$ into the same direction). The effect sizes for stratum 1 are chosen as those observed for the WHRadjBMI loci around STAB1, PPARG, or LYPLAL1. The power of [Diff_5e-8] is constant due to the lack of any filtering.