Would ascertainment bias affect the assumption of the random segregation of alleles?
Replying to “Would ascertainment bias affect the assumption of …”:
I would say no- ascertainment bias wouldn’t affect random segregation but it would affect the allele frequency of a SNP and the SNP effect on a trait within the study population, which could bias MR causal estimates. Having said that, I will let Dave to confirm this
Replying to “Would ascertainment bias affect the assumption of …”:
You need to consider other issues, like collider bias, if the outcome and exposure effect study selection
Replying to “Would ascertainment bias affect the assumption of …”:
I agree with both those responses. Ideally, the most powerful sort of MR study is a within family MR design where any segregation is observed within families. Most MR studies however are population based studies in unrelated individuals which technically are an approximation to the ideal case of a within family design. In population based studies, then ascertainment can become important potentially influencing the magnitude of the SNP-exposure and/or SNP-outcome associations. For more discussion, please see here: Within family Mendelian randomization studies - PubMed