June Virtual Seminar by Matt Robinson: an interview

Predicting diseases, one step ahead

Interview by Vivienne Baillie Gerritsen, Science writer at SIB

Our health depends not only on the genes that Nature dealt us at birth, but also on the environment we evolve in and the kind of life we lead. Smoking, stress, a lack of exercise or too much of it, pollution, overeating, undereating, irregular sleep, regular exposure to sun, a pint of beer after work and a glass of wine for dinner... all these conditions we live in, like the habits we acquire and indulge in routinely, can actually end up leaving their mark on our DNA.

Although these marks rarely affect us, they have proved to be important in predicting an individual’s state of health. “What I am interested in”, says Matt Robinson – SIB Group Leader at the Department of Computational Biology at the University of Lausanne – “is designing models that can predict the risk of a disease in a person and for this, you have to be able to take into account many different environmental, clinical, genetic and epigenetic measurements that are becoming increasingly available.”

This is the realm of complex trait genetics. Common conditions such as diabetes and coronary artery disease but also our height and body mass index (BMI), for example, are caused by multiple genetic and environmental factors. If you can find a way to identify these factors on a person’s genome – or parts of their genome – then measures can be taken to circumvent the oncome of a possible disease by way of early diagnosis and treatment, and perhaps even a change in lifestyle. What exactly, though, are these markers that taint our genome? “Chemical modifications of parts of our DNA, like methylation for example, which reflect the lifestyle we live”, explains Robinson.

Robinson and his team led their research on large European population samples, using a Bayesian learning approach to design their models. These models examined the link between genetic and epigenetic factors (SNPs and methylation) and common complex diseases. Their results? “The Bayesian model we have designed has proved to be a good one and, so far, outperforms any other methods” says Robinson. “More importantly perhaps, it demonstrates how by taking a simple blood test, not only can we learn a lot more about an individual’s disease risk but our understanding of the underlying processes involved is also heightened.”


After earning his PhD in Quantitative Genetics at the University of Edinburgh (UK) in 2008, Robinson pursued his career with a first post-doctoral training at the Department of Animal and Plant Science at the University of Sheffield (UK), and a second at the Department of Animal Ecology at Uppsala University in Sweden. On his return at the University of Sheffield in 2010, he took up a NERC Early Career fellowship in the same department he had left a year earlier. Robinson then shifted to human medical genetics and became group leader in 2013 at the Institute of Molecular Biosciences of the University of Queensland in Australia. He came back to Europe in 2017 and is currently Assistant Professor at the Complex Trait Genetics Group of the Department of Computational Biology at the University of Lausanne.

More on our Virtual Seminar Series