Research

Research Focus

Our group focuses on dissecting the genetic basis of common complex traits and diseases and the development of statistical genetics methods and software to accomplish this task. We work across a range of different phenotypes and diseases including osteoporosis, autoimmune diseases including ankylosing spondylitis and multiple sclerosis, overwhelming bacterial infection (sepsis), laterality (handedness), diet and cardio-metabolic disease, and early life phenotypes such as birth weight. We are active in statistical genetics methods development including approaches for gene mapping, variance decomposition and causal modelling. We have a particular interest in Mendelian randomization (MR) and structural equation modelling (SEM), and using these methods to investigate the Developmental Origins of Health and Disease (DOHaD) and other causal questions in genetic epidemiology. We have a lot of large scale genetics datasets in house including the UK Biobank study, and have excellent collaborative relationships with a number of the world’s premier cohort studies including ALSPAC, MOBA and HUNT studies.

Research Projects

We are always looking for bright, self-motivated and computationally literate students to join our group and we accept PhD and honours/masters students from a wide range of backgrounds including genetics, epidemiology, computer science, mathematics/statistics, engineering and psychology. In general applicants should have a high degree of numerical literacy and be comfortable/have a desire to work with large datasets. Applicants should feel free to contact David Evans at email d.evans1@uq.edu.au or other members of the Statistical Genetics Lab to discuss potential projects.

The following are a (non-exhaustive) list of projects currently being offered at the PhD level. Elements of the below may also be suitable for stand-alone honours projects:

Elucidating the Genetic and Genomic Basis of Septic Shock: Sepsis is an abnormal host response to infection, which results in organ and tissue damage, and in severe cases death. The most severe form of sepsis- septic shock- has a mortality rate around 30%. We have one of the largest –omics cohorts of septic shock in the world located in house. We have GWAS, RNA Seq, methylation, NMR metabolomics data and detailed clinical phenotypes on up to 600 individuals. We are interested in examining the relationship between GWAS, gene expression, methylation and metabolomic quantities and clinical variables related to recovery/mortality from septic shock.

Addressing Pleiotropy in Mendelian randomization Studies: Mendelian randomization (MR) is an epidemiological method that uses genetic variants robustly associated with modifiable environmental exposures to investigate the causal relationship between these environmental exposures and risk of disease. Genetic pleiotropy (i.e. the genetic variants influencing >1 phenotype) is one of the biggest threats to the validity of such studies. In this project we will investigate a new MR procedure we have developed that we believe is robust to many forms of genetic pleiotropy and assess its statistical properties and performance on real data from the UK Biobank and other large scale genetics cohorts.

Using Mendelian randomization to Investigate the Developmental Origins of Health and Disease: There is a well-documented observational relationship between low birthweight infants and increased risk of cardiometabolic disease in later life (e.g. type 2 diabetes, hypertension, cardiovascular disease). This inverse association was initially interpreted as resulting from developmental compensations to an adverse intrauterine environment, which in turn led to long-term changes to offspring physiology and increased susceptibility to cardiometabolic disease. This theory was christened the “Developmental Origins of Health and Disease” (DOHAD) and has been one of the preeminent paradigms in life-course epidemiology over the last thirty years.

In a series of recent ground-breaking studies (Horikoshi et al. 2016 Nature; Warrington et al. 2019 Nat Genet), we showed that the inverse correlation between birthweight and cardiometabolic disease may in fact be predominantly mediated by genetic rather than environmental factors. However, maternal and offspring genotypes are correlated, meaning that dissecting the genetic and environmental contributions to this relationship is fraught with interpretational difficulties, including the possibility that any genetic effects may be mediated through the mother’s (rather than the offspring’s) genotype operating on the intrauterine environment. The aim of this project is to develop and apply novel statistical genetics methods to large genetic datasets to investigate the Developmental Origins of Health and Disease (DOHAD).