In the quest for causality: Mendelian Randomisation Conference 2017

Observational study designs are often so rife with biases, it makes studying causal associations next to impossible. Randomised controlled trials are the gold standard but expensive, and at times, not feasible, or run the risk of being unethical. Enter large-scale accessible genomics and big data analysis: a novel approach to infer causal associations, known as Mendelian randomisation (MR).

Genetic variants for specific traits offer a relatively unbiased approach to validate or refute epidemiological associations that is robust from confounding and reverse causality. The approach exploits the random distribution of alleles (genes for specific traits like hair colour) during meiosis (an event that distributes genetic information from both parents to create a unique person).

These genetic variants are increasingly being explored and catalogued by large-scale genome-wide association studies. MR-based study designs extract data from these sources to conduct instrumental variable (IV) analysis – where the genetic variants proxy specific traits (such as levels of a biomarker) and causal associations with endpoints of interest (such as heart disease) are then studied.

However, it also brings with it a whole host of methodological issues that needs to be scrutinised and peer-reviewed rigorously. This 2nd landmark Mendelian Randomization Conference in the beautiful city of Bristol – known as ‘UK’s Green Capital’ – attempts to do just that. It brought together researchers from all over the world, in the hopes of fostering collaborations. It revealed a network of investigators dedicated to finding novel drug targets, creating innovative MR-based methods, probing epidemiological associations and even extending these to more complicated topics like foetal origins hypothesis.

This year the conference was held at the iconic Colston Hall which has been a classical music venue for more than a hundred years. It was an immersive three-day experience listening to the symphonies of renowned epidemiologists and statisticians, brilliantly orchestrated by the MRC Integrative Epidemiology Unit (MRCIEU) of the University of Bristol.

The conference opened up with a keynote lecture from Professor Martijn Katan, who was the first person to propose using genetic variants to study causal associations. This was immediately followed by brief overviews by principal investigators (PIs) of large-scale biobanks making MR-based studies possible. The first two PIs – Professor Rory Collins and Dr Zhengming Chen – are both based in the department. They showcased the sheer scale of their respective biobanks, not just in terms of size (half a million participants each!) but the outstanding breadth and depth of phenotypes available for researchers to study.

Since starting to develop in the early 2000s, MR studies have steadily risen in prominence with papers published by researchers ranging from statisticians to social scientists to drug target researchers.

Growth of the number of MR studies as estimated by a PubMed search of “Mendelian Randomisation” or “Mendelian Radomization” on the 7th December 2015 (US National Library of Medicine).

Source: www.authorea.com/users/159976/articles/178917-mendelian-randomisation-a-minireview

Indeed, the three-day event profile was curated thematically to be able to accommodate 50+ talks in such a short span of time. This meant we had to prioritise what talks we would want to attend; particularly troubling for young novice researchers like us who’d be interested in listening to everything!

All in all, it was an exciting three days packed with rich content that had us wanting for more. Being a first year DPhil student with nothing to really show for it yet, I was enthralled by the quality and volume of research already done in this field. My research focuses on finding druggable inflammatory targets for cardiovascular disease and diabetes, and the conference left me with more questions than answers. Nonetheless, the opportunity to experience an event of this scale first-hand made me grateful for how far I’ve come and, more importantly, how far I have yet to go.

Mohd Karim was awarded the Oxford Center for Islamic Studies Scholarship (OxCIS) to study for an MSc in Global Health Science (2015-2016) in the department (Linacre College). Soon after, he was also nominated for the Medical Science Graduate School studentship award (2016-) to pursue his DPhil in Population Health at CTSU (St Edmunds Hall). His current work is on the China Kadoorie Biobank, exploring causal associations of inflammatory biomarkers (CRP, IL-1,6, TNF-alpha etc) with risk of cardiovascular disease, using a Mendelian randomisation approach.