Impulsive aggression is a key feature of a number of psychiatric disorders including schizophrenia, bipolar disorder, post-traumatic stress disorder, attention deficit hyperactivity disorder, and conduct disorder. It has a deep impact on the patient, as well as society. Acts of aggression account for 1.5 million deaths annually worldwide, yet effective treatments for impulsive aggression are very limited. A better understanding of the development of aggressive behavior and the underlying neural circuits will allow for the development of interventions and targeted treatments for excessive aggression found in psychiatric disorders. From recent basic science research, using a novel transgenic mouse model we recently identified a developmental sensitive period for the effect of 5-HT1BR on aggression. Specifically, alterations in 5-HT1BR signaling during the peri-pubertal period lead to increased impulsive aggression in mice throughout life. Now, through epidemiological analysis of longitudinal studies, we will investigate the presence of a similar sensitive period for the development of aggression in humans.

Eczema is a complex disease, with heterogeneous presentations, clinical courses and outcomes. Despite this, our approach to characterising eczema is surprisingly unidimensional. Failure to capture disease complexity when characterising patients is a major stumbling block, which limits understanding of eczema aetiology. It also hinders the practice of stratified medicine. At present, we are failing our patients, with inadequate strategies and therapeutic interventions. Novel therapeutic agents for eczema are on the horizon; multidimensional phenotyping would allow us to understand how to better use these and existing therapies to help our patients, by allowing us to tailor treatments for individual patients (personalized medicine). Understanding phenotypes would also help inform prevention strategies, which are important giving the rising prevalence of eczema.
We will use novel bioinformatic and statistical approaches to identify specific groups of individuals with eczema (phenotypes) in the Avon Longitudinal Study of Parents and Children (ALSPAC) prospective birth cohort study. Phenotypes will combine a variety of multidimensional characteristics including characteristics of the participant’s eczema and clinical history, immunological, genetic and novel biomarker characteristics. We will then assess predictors of clinical phenotypes.

The epigenome is considered by some as the key missing piece to understanding the etiology of many biomedical diseases (Mill & Heijmans, Nat Rev Gen 2013). Epigenetic modifications are non-sequence based alterations to DNA that can induce stable changes in the regulation of gene expression. Technological advances in high-throughput DNA analysis have facilitated genome-wide examination of epigenetic modifications and large-scale association testing with disease-related outcomes. Such studies have found epigenetic markers to be associated with a variety of diseases (such as cancer and diabetes) as well as with environmental exposures (including BMI, smoking, alcohol and diet) (Relton & Davey Smith, PLoS Med 2010). Furthermore, epigenetic variation can contribute to the development of a disease or be a consequence of it. Although distinguishing the direction of causation represents a major challenge for epigenetic studies, epigenome-wide association studies (EWAS) can contribute novel molecular insights into the genetic and environmental determinants of disease-related phenotypes (Relton & Davey Smith).

We propose to conduct an EWAS exploring epigenetic markers associated with fatty acids (FAs). Polyunsaturated FAs are involved in key biological processes, including inflammatory responses, gene expression and cellular fluidity (Nakamura et al. Ann Rev Nutr 2004). Omega-3 and Omega-6 are the two main classes of polyunsaturated FAs and imbalance in the ratio of omega-6 to Omega-3 FAs and/or deficiency in omega-3 FA has been associated with adverse outcomes such as cardiovascular disease and diabetes (Yashodhara et al. J Postgrad Med 2009). Assessing epigenomic variation related to FAs may uncover novel biological insights into the role of FAs in health and disease.

Additionally we would like to utilize SNPs either related to fatty acids or methylation, to perform bidirectional MR analyses to tease out any possible causal relationships underlying methylation-fatty acid associations.

Anorexia nerovsa is a severe mental illness most often affecting adolescent girls. The etiology of the disorder is complex and involves genetics as well as environmental aspects. We have shown that epigenetic factors can contribute to the disorder. In the current Project, we have already measured epigenetic marks (i.e. DNA methylation) on a genome-wide level in girls participating in a psychotherapy-trial for anorexia nervosa. We want to compare our findings to the data from girls participating in the ALSPAC study that show no signs of an eating disorder.

Complex diseases are typically caused by a mixture of genetic, environmental and epigenetic effects. A number of prenatal risk factors, including intra-uterine growth restriction (IUGR) and gestational diabetes mellitus (GDM), have been shown to determine signatures in the blood methylome. However, most studies on this topic fall short are underpowered, or fail to take genetic effects into account to investigate causality. Birth cohorts with multiple tissue samples and deep phenotyping offer an opportunity to investigate the interplay of genetics, epigenetics and the environment affecting foetal and child growth. Moreover, my project aims to specifically determine which changes in DNA methylation are causal to subsequent disease states by combining genetic and epigenetic data in Mendelian Randomization experiments. The discovered results can then be used to develop biological tools to detect children at risk stratify risk groups and develop prevention strategies.

The work I propose to undertake follows on from my recent research in cross-sectional epigenome-wide association studies (EWAS) for T2D and obesity In adults (Wahl*, Drong* et al. under review). However, I have found previously that most of the strongest signals of methylations associations display a reverse causal relationship. I am thus interested to investigate the influence of early-life events, such as foetal growth and gestational diabetes (GDM) on patterns of DNA methylation. To achieve this, I aim to employ quantitative skills to develop robust methodology to take into effect confounding effects from experimental confounders (mixture of foetal/maternal tissues), maternal/paternal genotypes and to develop a robust, reusable pipeline for Mendelian Randomization in birth cohorts. Ultimately, I aim to link epigenetic markers both identified for GDM and IUGR with future health outcomes, and develop biomarkers for the effects.

Firstly, I aim to apply my analysis methodology to utilise the rich data sets curated by the proposed research sponsors to detect associations of DNA methylation with a number of phenotypes. I will perform a large scale EWAS case/control studies for GDM. Secondly, I will lead for the analysis of epigenetic data from biological samples for IGUR in the INTERBIO-21 study. This includes development of the experimental designs for large-scale epigenome-wide association scans. Lastly, I will be utilizing genetic variants as instrumental variables in two-step Mendelian Randomization to determine whether epigenetic markers are in causal pathways linking intra-uterine events to the child's phenotypes. Thus I will apply an informed approach about causality to separately detect genetic associations to avoid bias and will be able to detect maternal genetic and epigenetic confounding.

This will allow me to be in a unique position by extending the scope of simple cross-sectional EWAS to include not only causal analysis, but also a robust characterisation with respect to biological and technical confounders.
By focussing my hypothesis on causal pathways, my work can filter out reverse-causal confounders facilitate personalised prevention and novel drug target discovery. If successful, my research will provide accurate tools to guide childhood interventions through early biomarkers of the intrauterine environment.

ALSPAC data related to body size and blood pressure will be used to produce a new dataset with similar characteristics, but which is entirely simulated and contains none of the original ALSPAC data.

The simulated dataset will be shared with collaborators from two Virtual Reality (VR) Games Companies, who are taking part in a data visualisation competition organised by the Wellcome Trust. ALSPAC was invited to take part in the competition, by Wellcome Trust, because it is an excellent example of 'big health data'.

No ALSPAC data will be shared with the Games Companies.

There exists a persistent and strikingly large correlation between education and health. For example, in the Netherlands, the higher educated can expect to live 20 years longer in good health compared with individuals who finished only primary school. Despite the stark correlation, studies carefully addressing causality suggest that the causal effect of education on health is much smaller than the correlation suggests (Van Kippersluis et al. 2011; Meghir et al. 2012; Clark and Royer, 2013). This apparent contradiction can be explained either by a strong effect of health on educational attainment, or alternatively by the same set of individual characteristics (e.g. time preference, cognitive ability) strongly affecting both health and educational attainment.

A recent series of papers by Conti and Heckman (2010), Conti et al. (2010), and Bijwaard et al. (2015) use birth cohort data with an extensive set of early childhood characteristics to disentangle the effect of (i) education and (ii) early childhood cognitive and non-cognitive skills, on health outcomes. The results show that for most health outcomes at least half of the association between education and health is driven by cognitive and non-cognitive abilities and early childhood social background.

The set of characteristics available in the birth cohorts used is comprehensive, and this series of papers has substantially improved our understanding of the association between education and health. However, the list of characteristics is not exhaustive. In particular, two potentially important channels are missing: (1) the effect of health on educational attainment is measured poorly by imperfect proxies such as childhood height, or is even entirely absent, and (2) the role of genes is not studied at all. That is, education and health may share a common genetic architecture, in which the same set of genes influence both education and health outcomes (biological pleiotropy). This is the hypothesis we intend to test in this project.

Recently, Boardman et al. (2014) have shown that indeed there is genetic correlation between education and health. However, as these authors rightly point out, this correlation does not shed light on the underlying mechanism. The genetic correlation does not imply biological pleiotropy, but could also occur due to mediation pleiotropy (for example, genetic factors influencing health, and in turn health influencing educational attainment).

Biological pleiotropy is best studied on the level of individual genetic variants (Solovieff et al., 2013). However, current methods seem not able to provide definitive answers here, because individual genetic variants have only limited explanatory power. Therefore, we propose two strategies to improve our understanding of the relation between health and education.

References:
Bijwaard, G.E., Van Kippersluis, H., & Veenman, J. (2015). “Education and health: The role of cognitive ability”, Journal of Health Economics, 42: 29-43.
Boardman, J.D., Domingue, B.W., & Daw, J. (2014). “What can genes tell us about the relationship between education and health?”, Social Science & Medicine, doi:10.1016/j.socscimed.2014.08.001.
Clark, D., & Royer, H. (2013). “The effect of education on adult mortality and health: Evidence from Britain”, American Economic Review, 103: 2087-2120.
Conti, G., & Heckman, J.J. (2010). “Understanding the early origins of the education-health gradient: A framework that can also be applied to analyze gene-environment interactions”, Perspectives on Psychological Science, 5: 585-605.
Conti, G., Heckman, J.J., & Urzua, S. (2010). “The education-health gradient”, American Economic Review Papers and Proceedings, 100: 234-238.
Meghir, C., Palme, M., & Simeonova, E. (2013). “Education, cognition and health: Evidence from a social experiment”, NBER Working Paper 19002.
Solovieff, N., Cotsapas, C., Lee, P.H., Purcell, S.M., & Smoller, J.W. (2013). “Pleiotropy in complex traits: challenges and strategies”, Nature Reviews Genetics, 14: 483-495.
Van Kippersluis, H., O'Donnell, O., & Van Doorslaer, E. (2011). Long run returns to education: Does schooling lead to an extended old age?", Journal of Human Resources, 46: 695-721.

Urogenital malformations in male neonates is thought to be associated with exposure of the foetus to mixtures of endocrine disrupting chemicals during pregnancy. However, the causative chemical agents and their mixtures have not yet been identified. This project will use new discovery-based chemical methods to analyse samples of maternal urine from an ALSPAC cohort containing cases with urogenital malformations. The goal of the work is to identify key mixtures of chemical contaminants that have endocrine disrupting activity and cause malformations in the developing foetus.

The number of children who are affected by parental alcohol misuse is largely unknown although estimates suggest a third of all UK children live with at least one parent who uses alcohol hazardously. How this impacts on their health, mental health and education is unclear.