Steiger Test avoids reverse causation
The following steiger test method is a new methodology. After adding it, it improves the robustness of extracting instrumental variables. Otherwise, it will not affect the content of the article.
Depend on your own needs. During the analysis process, if the GWAS data does not have a samplesize and you need this content, please increase the samplesize yourself before proceeding.
In the Mendelian randomization causality analysis of exposure A->outcome B, the correlation between SNP and exposure A is greater than that of outcome B. Otherwise, reverse causality will occur, that is, SNP cannot be used as an instrumental variable for exposure (Mendelian Three major assumptions of exclusivity)
In the process of extracting the instrumental variables of exposure, in addition to filtering SNPs that are significantly associated with the outcome (default 5e-5 in the code)
Steiger Test statistics can be introduced to filter out this part of SNPs
The role of Steiger Test:
1. Test whether the correlation between SNP and outcome is greater than the exposure
2. SNPs that fail this test may not be related to exposure and should be screened before analysis
Code usage:
r.test from psych package
Upgrade MendelR to version 6.0