贝叶斯共定位分析指南：如何评估MR工具变量与暴露的因果共享？

For significant results, we used the Bayesian colocalization assessment tool to evaluate whether there are shared causal variants between the instrument variable and the exposure. Using the SNPs that passed the screening in the MR analysis as anchors, we extracted SNPs within a 500 kb region upstream and downstream of each anchor for the analysis. This analysis is based on five hypotheses: H0 assumes no association for either phenotype; H1 assumes an association only with the exposure; H2 assumes an association only with the outcome; H3 assumes an association with both but driven by different causal variants; H4 assumes an association with both and sharing the same causal variant. According to established standards from previous studies, a posterior probability (PP4) ≥ 0.80 for H4 indicates strong evidence of colocalization, while 0.50 ≤ PP4 < 0.80 is considered moderate evidence.