Table 1. Simulation results under different relationships between the genetic variants and the exposure at each time point.

		MR	MVMR
Exposures associated with different liability periods
β 1	Liability effect	0.344	0.200
	Effect estimate	0.340	0.1958
	Est. Std. Error	0.029	0.0107
	Simulation Std. Error	0.011	0.0106
	Absolute bias	0.010	0.0092
	Coverage	100%	93%
	F-statistic	96.31
	Conditional F-statistic		55.76
	No. SNPs	72	114
β 2	Liability effect	0.376	0.300
	Effect estimate	0.371	0.297
	Est. Std. Error	0.015	0.009
	Simulation Std. Error	0.008	0.009
	Absolute bias	0.008	0.008
	Coverage	99%	94%
	F-statistic	129.31
	Conditional F-statistic		78.01
	No. SNPs	83	114
Exposures associated with the same liability period
β 1	Liability effect	0.530	0.200
	Effect estimate	0.519	0.207
	Est. Std. Error	0.011	0.080
	Simulation Std. Error	0.011	0.080
	Absolute bias	0.013	0.063
	Coverage	82%	94%
	F-statistic	96.31
	Conditional F-statistic		1.06
	No. SNPs	72	86
β 2	Liability effect	0.480	0.300
	Effect estimate	0.474	0.288
	Est. Std. Error	0.009	0.073
	Simulation Std. Error	0.009	0.072
	Absolute bias	0.009	0.058
	Coverage	89%	94%
	F-statistic	115.76
	Conditional F-statistic		1.06
	No. SNPs	83	86

N = 150,000, reps = 2000, Direct effect of exposures on the outcome; X₁: β₁ = 0.2, X₂: β₂ = 0.3. Effect of X₁ on X₂ = 0.1. Effect of the liability for each exposure point are given in the table. Absolute bias is the mean value of the absolute bias of the effect estimate across the simulations. For each of Effect estimate, Est. Std Error, F-statistic and Conditional F-statistic mean values across each iteration of the simulation are reported. Simulation Std. Error is the estimated standard error in the effect estimate across the repetitions in the simulation. Coverage gives the proportion of times the true effect estimate falls within the 95%CI, No. SNPs is the mean number of SNPS selected for estimation.