Figure 1. Flowchart of univariable and multivariable MR analyses.

(A) Univariable MR. MR has three fundamental assumptions. (1) Relevance assumption: the genetic variants (instrumental variables) must be strongly correlated with exposure(s) (P < 5 × 10⁻⁸) (r² < 0.001 and distance > 10 000 kb, the SNPs in pairwise linkage disequilibrium). (2) Independence assumption: no unmeasured confounders of the correlations existed between genetic variants and outcome(s). (3) Exclusion restriction assumption: the genetic variants influenced the outcome(s) only via exposure(s). (B) Multivariable MR: MR, Mendelian randomization; educational attainment, years of schooling; BMI, body mass index; smoking behavior, smoking behavior (never vs. initiation); using mobile phone, length of mobile phone use; watching TV, time spent watching television; and SNP, single nucleotide polymorphism.