Using Mendelian Randomization to infer causality in life course epidemiology

Rationale
Observational associations in epidemiology may suffer from residual confounding, even after adjustment for measured variables.
Mendelian Randomization (MR) is a tool used to improve the strength of causal inference from observational data. In MR, genetic variants related to the exposure of interest are used to give a causal, unconfounded estimate of the exposure-outcome association. MR has successfully been applied to many research questions in epidemiology. However, to date, applications have focused on the situation where there is a single measure of the exposure and a single measure of the outcome.
In life course epidemiology, the focus is on assessing the influence of exposures throughout the life course on later health and disease; this often requires the analysis of longitudinal data. Using appropriate longitudinal analysis methods can help to understand these relationships, but applying MR to such questions would strengthen causal inference.

Aims & Objectives
The aim of this PhD is to extend the framework of MR to situations where either the exposure of interest, the outcome, or both are measured repeatedly.

Depending on the interests and strengths of the student, this PhD could be an epidemiological PhD, applying MR to problems of interest, or it could be a more statistical PhD, developing MR methodology, or it could be a combination of epidemiology and statistics.

Methods
Depending on your interests, you could focus on various areas; some examples are:
1. A single measure of the exposure, repeated measures of an outcome:
Observational data suggests that maternal smoking is associated with blood pressure during pregnancy. MR using longitudinal models of maternal blood pressure over pregnancy as the outcome could help to understand whether the association changes in different periods of pregnancy.
2. Repeated measures of the exposure, a single measure of the outcome:
Evidence from MR studies suggests that height has a causal effect on cognitive and behavioural outcomes. This analysis could be extended to determine whether height growth in specific periods represents a sensitive period in the development of the outcomes
3. Repeated measures of both exposure and outcome:
MR studies have attempted to untangle the causality of observed associations between obesity and depression. MR using trajectories of fatness and mental health could help to untangle this.