Using Mendelian Randomisation to investigate the associations between dietary exposures and risk of ovarian and endometrial cancer

Suitable for a Masters (preferably part-time) or Honours student. Some experience in biostatistics and data analysis is essential and a background in epidemiology and/or an interest in cancer are highly desirable.

Endometrial cancer and ovarian cancer are the most common gynaecological cancers other than cervical cancer among women in high-income countries. Observational studies have evaluated some potentially modifiable factors including physical activity and some dietary components (e.g. coffee, mono-unsaturated fatty acid, calcium, vitamin B/C, iron, selenium and zinc) in the causation and survival of these gynaecological cancers, but data have been either limited or inconsistent. However, observational studies may not provide robust evidence of causality as it is often vulnerable towards confounding bias, reverse causation and/or measurement error. Mendelian randomization (MR) is an analytic approach that can mitigate these problems using instrumental variables constructed from genetic variants to provide insights into causality. Subject to specific mathematical assumptions, analyses using germline genetic variants as instrumental variables are less susceptible to biases from confounding and reverse causation as the allocation of these germline variants cannot be influenced by later-year lifestyle and environmental factors. This property allows specific risk factors/exposures to be proxied via genetic predisposition (e.g. genetically predicted coffee intake can be estimated through the aggregation of alleles associated with increased daily coffee intake in large genetic studies) which can be used to infer causality between risk factors (e.g. coffee intake) and the outcome. Findings from MR can help re-prioritise resources toward trials of the most promising interventions. This project will apply various MR techniques in large population-based studies and biobanks to investigate the relationship between potentially modifiable factors and ovarian and endometrial cancer outcomes.