Discovery in ARIC

Proteome-wide MR approach was conducted to identify potential therapeutic targets for various heart failure (HF) outcomes. For proteins with only one single nucleotide polymorphism (SNP) available as the genetic instrument, the Wald ratio was calculated. For proteins with two or more SNPs as genetic instruments, the inverse variance weighted (IVW) was performed as the primary method [35]. In this study, we employed the Bonferroni correction for multiple testing, implementing an allowable type I error rate (α) of 0.05/N, where N indicates the number of outcomes tested under study. Supplementary methods, including weighted median and MR-Egger, were employed for proteins with three or more SNPs as genetic instruments. Other sensitivity analyses, such as MR- Egger regression intercept analysis and Cochran’s Q statistical analysis, were performed to evaluate horizontal pleiotropy and heterogeneity, ensuring the robustness of the results [36–38].

Colocalization analysis was utilized to further validate the findings. The colocalization method primarily assessed the validity of the instrumental variable hypothesis, taking into account whether the potential genetic association between exposure and outcome was explained by linkage disequilibrium [39–41]. The colocalization analysis yielded several posterior probability hypotheses (PPH) [41]:(1)No genetic association between the two traits (PPH0); (2) Only trait 1 had a genetic association (PPH1); (3) Only trait 2 had a genetic association (PPH2); (4) The two traits were correlated but had different causal variables (PPH3); (5) The two traits shared a causal variant, with the first trait influencing the second trait (PPH4). In this study, we conducted colocalization analysis using all SNP data within the entire cis-coding region of pQTL (± 500Kb). To satisfy the causal hypothesis, PPH4 commonly required a threshold of 80% [42].