Advanced Mendelian Randomization

A live online short course from Bristol Medical School for researchers already running Mendelian randomization (MR) studies, covering the latest methods and how to interpret them, taught by researchers at the MRC Integrative Epidemiology Unit.

Mendelian randomization is an approach to causal inference that uses genetic variants as instrumental variables to estimate causal effects of a risk factor on disease or health outcomes. Methods within the field and understanding of the strengths and limitations of Mendelian randomization are developing rapidly and academics working in the MRC Integrative Epidemiology Unit and the University of Bristol, including the tutors on this course, are at the forefront of this development.

The aim of this course is to provide participants with an advanced understanding of the methods available for Mendelian Randomization including estimation, limitations and interpretation of results.

For dementia researchers, MR is increasingly used to test whether risk factors such as blood pressure, education or lipids play a causal role in dementia, drawing on large genetics datasets. If you work in dementia epidemiology or biostatistics, this course takes you beyond the core methods.

What the course covers

1. methods for pleiotropy robust MR;
2. methods and interpretation of multivariable MR;
3. interpretation of MR with time-varying exposures;
4. methods for within-family MR;
5. MR with multi-ancestry data;
6. MR with a single SNP as an instrument;
7. the implications of selection bias and collider bias for MR;
8. practical application of pleiotropy robust methods, multivariable MR and within-family MR; and
9. opportunities for wider discussion about MR with tutors on the course.

What you will learn

By the end of the course you should be able to:

• Interpret an MR study, including its assumptions and limitations
• Explain advanced pleiotropy-robust estimation methods
• Conduct and interpret multivariable MR
• Interpret MR studies with time-varying exposures
• Understand the potential and the limits of MR with multi-ancestry data
• Interpret and conduct within-family MR
• Explain methods for, and the limits of, MR with a single SNP as the instrument
• Explain how collider and selection bias affect MR

Tutors

The course is led by Dr Eleanor Sanderson, Dr Kaitlin Wade and Dr Apostolos Gkatzionis, all researchers at the MRC Integrative Epidemiology Unit, University of Bristol.

Who should attend

The course is open to all, subject to the prerequisites, and is aimed at researchers already conducting MR studies who want to get to grips with recent methodological developments and how to interpret them.

You should have completed the Bristol Mendelian Randomization short course or equivalent, and have experience of running MR analyses. The practicals use R, so you need a working understanding of R, though not of MR estimation in R specifically. You will need a recent version of R, or you can use the cloud-based Posit Cloud instead.

Format

The course runs online over two and a half days, Wednesday 25 to Friday 27 November 2026, with live lectures, practicals and discussion and Q&A. Course materials go up on Blackboard Ultra one to two weeks beforehand. Live lectures are recorded and uploaded within 24 hours, and you keep access for five months after the course. Breakout sessions are not recorded.

Cost

The fee is £630. Bookings close two weeks before the course starts.

Book a place

Organised by the Bristol Medical School Short Course Programme, University of Bristol. Full details on the course page.