Last week, I was able to gather all of the data needed for analysis. This week, I studied the documentation for TwoSampleMR and practiced using Mendelian Randomization to assess the relationship between body mass index (BMI) and Coronary Heart Disease (CHD). The guide I followed for this can be found here.
To do so, I first used the BMI GWAS to locate the single nucleotide polymorphisms (SNPs) associated with BMI. Then, I searched the CHD GWAS for those same SNPs and extracted information associated with them, including each beta-value, standard error, affected allele, reference allele, etc. Next, I used the harmonization feature of TwoSampleMR to ensure that genotypes were coded consistently between the two sets of summary data. With all the data harmonized, I then used the inverse variance weighted method from TwoSampleMR to estimate the effect of BMI (the exposure) on CHD (the outcome).
I also learned about creating graphs, including forest and funnel plots. These help illustrate the effects of each SNP and will be of great importance in my actual analysis, which will begin next week!
Thanks for reading!