Some estimates in the simulation are highly biased. Is it a sign of a causal gap because the investigated SNP is "linked" to another SNP in the simulation?
In order to sort it out, for each biased estimate we can:
[x] Check whether the investigated SNPs are associated with other SNPs in the simulation
[x] Rerun the estimation, adjusting for highly associated SNPs and see if the bias goes away.
If this turns out to be true, we need to adjust the simulation. For instance we can:
[x] Fit the propensity score models for each variant for interest (or all variants in the simulation? Only do if necessary => I don't think it should be the case but who knows...)
[x] Regenerate data by sampling variants from this model
Some estimates in the simulation are highly biased. Is it a sign of a causal gap because the investigated SNP is "linked" to another SNP in the simulation?
In order to sort it out, for each biased estimate we can:
If this turns out to be true, we need to adjust the simulation. For instance we can:
Hopefully this will fix the bias.