Our Genetics in Drug Development course is the best place to learn about the power of human genetics for aiding with the discovery and development of therapeutic targets. From now on, we will be operating the course on a continuous recruitment basis. The course is a mixture of pre-recorded talks, computer practicals, and live sessions to help with coding and [Read more...]

This blog post is an addendum to the commentary "Towards more reliable non-linear mendelian randomization investigations" to provide practical advice on the implementation of non-linear Mendelian randomization.  Non-linear Mendelian randomization is an extension to standard Mendelian randomization that aims to understand the shape of the causal relationship between an exposure and outcome using genetic variants as instrumental variables. As this [Read more...]

The MendelianRandomization software package enables the implementation of various methods for Mendelian randomization using summarized data. This package has recently been updated a number of times to add extra functionality, and is now on version 0.9.0. Some references: MendelianRandomization: an R package for performing Mendelian randomization analyses using summarized data. This describes the basic operation of the package and core [Read more...]

Non-linear Mendelian randomization (or stratified Mendelian randomization) is an extension to standard Mendelian randomization that performs separate instrumental variable analyses in strata of the population. The motivation is to understand the causal dose—response curve relating the exposure to the outcome – how does the outcome vary on average when we intervene on the exposure at different levels of the exposure? [Read more...]

When performing a Mendelian randomization analysis based on a single gene region, it is possible to include only a single variant in the analysis. However, if multiple variants explain independent variance in the exposure, then the analysis will be more efficient using all those variants - even if they are partially correlated. With summarized data, this can be achieved using the [Read more...]

One question I’ve been asked a large number of times is how to perform a power calculation in Mendelian randomization for a binary exposure. We wrote a longer paper on the difficulties of MR with a binary exposure previously (Mendelian randomization with a binary exposure variable: interpretation and presentation of causal estimates), but we removed details on power calculation due to [Read more...]