Fig. 2: Simulation results in scenario 1 with autoregressive LD pattern.
From: A robust cis-Mendelian randomization method with application to drug target discovery
All 10 IVs have an effect on the exposure. A K1 = 0 (no invalid IV). B K1 = 4 invalid IVs. From top to bottom are empirical type-I error under the null, power under the alternative, and root mean squared error (RMSE) under the null. Colors represent different methods being compared. Source data are provided as a Source Data file.
