identify potential functionally relevant genes at the loci identified in GWAS and to identify novel trait-associated genes for five complex traits [22], and further applied the method to the analyses of another 28 traits [23]. It has been shown that eQTL effect in blood can be a proxy for eQTL effect in most relevant tissues for various traits or diseases [22,24]. Mendelian randomization is an instrumental variable analysis approach that uses genetic variants as instrumental variables (for example, eQTLs) to test whether an exposure (for example, the expression level of a gene) has a causal effect on an outcome (for example, trait value or disease risk) [25,26]. In this study, the phenotypic trait is the outcome (Y), gene expression is the exposure (X), and the top cis-eQTL that is strongly associated with gene expression is used as the instrumental variable (Z). The previous study [22] suggested that there are three models consistent with a significant association from the SMR test using only a single genetic variant. These three models include causality (Z → X → Y), pleiotropy (Z → X and Z → Y) and linkage (Z1 → X, Z2 → Y, and Z1 and Z2 are two variants in
