Neurodevelopmental disorders (NDDs) are conditions involving perturbed brain development with manifestations ranging from specific to global impairment of developmental domains such as communication, social interaction, motor skills, cognition, activity and emotion. This includes autism spectrum disorder (ASD), attention-deficit hyperactivity disorder (ADHD), disorders of language, learning and motor functions, intellectual disabilities, schizophrenia and bipolar disorder. NDDs are highly prevalent in the population (~5% of children), typically emerge early in life, and are associated with substantial and long term disability.

The aim of this study to investigate the role of potentially modifiable risk factors for NDDs, using Mendelian Randomization (MR). Multiple pre-/perinatal risk factors are associated with NDDs, including nutritional deficiency (vitamin D deficiency, fatty acids), thyroid dysfunction, inflammation and autoimmune conditions, metabolic conditions, stress, smoking and low birth weight. However, results from epidemiological studies have been inconclusive and the causal role of these risk factors has not been established for most NDDs. Studies using methods that strengthen causal inference are important so that the role of these potentially modifiable risk factors can be determined.

In this project we plan to use MR analysis, whereby genetic variants robustly associated with the risk factors above are used as instrumental variables (IV) to examine whether they have a causal effect on NDDs (categorical disorders and dimensional traits). We will use two-sample MR in which the IV-risk factor and IV-outcome associations are obtained from non-overlapping samples (Burgess et al., 2015). First, published GWAS studies will be used to identify genetic variants robustly associated with the risk factors. Second, these genetic variants will be used to calculate a genetic score for each risk factor in an independent target sample (ALSPAC). Standard linear and logistic regression models will be used to estimate the effect of each risk factor on the NDD trait of interest in the ALSPAC target sample, using the genetic scores as an instrument.

We will also employ additional methods such as positive and negative control designs to strengthen causal inference (Lawlor, Tilling & Davey Smith, 2017).

References

Burgess S, Scott RA, Timpson NJ, Davey Smith G, Thompson SG, Consortium E-I (2015). Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors. European Journal of Epidemiology 30, 543–552.

Lawlor DA, Tilling K, Davey Smith G (2017). Triangulation in aetiological epidemiology. International journal of epidemiology. [Epub ahead of print]