The power of the environment: Environmental mediation of genetic liability

The consequences of childhood obesity remain one of the biggest public health burdens in the UK. 1 in 3 children are already overweight or obese by age 10 years. This is especially problematic because obesity has been linked to many health conditions, such as diabetes, cancer and depression. There is now substantive evidence that obesity runs in families, but genetics alone cannot explain the current rise of obesity in the population. So environmental factors have been proposed to be causal for this development. It is this dual determination, that remains paradoxical - how is it possible that obesity is explained by both genes and environments at the same time and what does this complexity mean for public health interventions?
The proposed research for this fellowship will investigate these issues. My aim is to bring together methodological tools from two fields: polygenic scores and mediation analyses. Polygenic scores are individual level scores that summarise the genetic liability for an illness or characteristic. Mediation analysis first establishes the association between a risk factor and a later health outcome, and then aims to establish how much of this association can be explained by a third variable (the mediator).
In the context of childhood obesity, we already know that genetic and environmental factors are important. However, it is not known how environmental factors might be protective of genetic liability. In this fellowship, my aim is to understand how environmental factors can mediate or moderate the effect of genetic liability (indicated by a polygenic score) for BMI.
My main questions are:
1. Do parental feeding practices, such as offering food to soothe, enhance or buffer the association between genetic liability and later childhood body size, and what could be achieved if we changed parental feeding strategies?
2. Does the child's physical activity mediate the association between childhood genetic liability and later body size, and what would happen if levels of physical activity increased?
3. Can maternal genetic risk influence the child's body size, even though the genetic liability are not passed on?
4. What is the association between genetic liability and the built environment of the family home, such as access to green spaces and public transport?

In order to address these questions, I will combine data sets, which already have all the information that I need. This makes this research very efficient and pragmatic. Due to the novelty, I am planning to compare two different methods to investigate mediation - one is the traditional more informal approach of structural equation modelling and the other is a more recently developed formal counterfactual approach. Comparing these different methods will be very useful because it can strengthen our confidence in the results, but also can help other researchers studying similarly complicated but ubiquitous questions.
These questions are of great importance to public health. Polygenic scores are readily available through commercial providers, but also have recently been proposed to be included into NHS standard practice. However, it is unknown how polygenic scores relate to environmental interventions and if these interventions can buffer genetic liability for obesity. Consequences of childhood obesity persists as one of the biggest health challenges, which can affect the entire family. This research has the potential to uncover some of the mechanisms underlying childhood obesity and highlight the pathways by which environmental factors can be protective.

Consequences of childhood obesity remains one of the greatest public health burdens. There is substantive evidence that obesity is highly heritable, but genetics cannot account for the recent rise of obesity in the population. Environmental factors have been proposed to be causal, but the mechanisms of gene-environment interplay are not clear. Specifically, the extent to which environmental factors can mediate the association between genetic risk and later disease. This fellowship will address these gaps in the literature by combining polygenic scores with mediation analysis. The main objectives are: (i) Estimating the mediating effects of parental feeding practices and child physical activity on the association polygenic score associated with BMI (PGS-BMI) and later childhood BMI. (ii) Comparing traditional structural equation modelling mediation methods with modern counterfactual methods, and the evaluating the differences of these methods when investigating the genetic-environmental aetiology of childhood obesity. (iii) Investigating the impact of non-transmitted maternal PGS-BMI on child BMI. (iv) Examining the interaction between child PGS-BMI and the build environment of the family home.
Data will be from readily available longitudinal cohorts; Avon Longitudinal study of Parents and Children (ALSPAC, N~8650) and Millennium Cohort Study (MCS, N~10000). Exposures: PGS-BMI will be derived for both cohorts from GWAS summary statistics. Intermediate mediators: Parental reports of feeding behaviours (ALSPAC, N~5000), accelerometer data at 7 (MCS, N~6500) and 11 years (ALSPAC, N~4700), Geographical information systems data based on postcodes of families. Outcomes: Objectively measured BMI and body composition measures at 11 (MCS, N~100000), 12 (ALSPAC, N~7000) and 14 years (ALSPAC, N~6500, MCS, N~10000). Results advance our understanding of the underlying mechanisms of childhood obesity. The novel application of PGS in mediation analysis will be beneficial for oth

Who will benefit from this research?
Researchers in the fields of childhood obesity, public health and quantitative genetics
Paediatricians and other professionals working with children, e.g. teachers, nurses and social workers
Families with children with obesity
Policy makers
The public

How will they benefit from this research?
This interdisciplinary project will benefit researchers from multiple areas: (i) childhood obesity, (ii) public health, (iii) applied statisticians and researchers aiming to investigate potential environmental mediation of genetic effects. Previous research has investigated the impact of genetic and environmental factors on the development of childhood obesity. However, less is known how genetic and environmental factors interact, specifically the extent to which environmental factors can mediate genetic liability for higher BMI during childhood. This overarching question is at the centre of the proposed research and produced results will be of importance to other researchers studying childhood obesity and public health. In addition, the proposed research project will combine methodological tools from statistical genetics (specifically polygenic risk scores) with counterfactual-based mediation analysis. This combination has been relatively unexplored. Even though, this project focuses on childhood obesity, the same analytic strategy could be applied to other complex health outcomes, influenced by genetic and environmental factors, such as depression or smoking. This could to lead to further statistical developments from statisticians as well as hypotheses generation for other health related researchers aiming to disentangle complex aetiologies.
The research will be beneficial for Paediatricians and other professionals working with children, e.g. teachers, nurses and social workers. Childhood obesity is very complex, and new research findings will help these practitioners, as well as teachers, nurses and social workers to communicate to families about the protective environmental steps they can take to support their children.
Similarly, families with children with obesity will benefit from this research. Understanding the complex development and how genetic and environmental factors interact in the development of child obesity, is crucial to empower families to support their children's health in the best way they can. In addition, emphasising the biological underpinnings of childhood obesity has the potential to reduce some of the blame and stigmatisation that is often connected to childhood obesity.
This project will be beneficial to public health and policy makers. The methodological approach suggested for this fellowship, provides the possibility to understand the potential effect environmental factors could have on the development of childhood obesity. Importantly, the selected environmental factors, parental feeding, physical activity and the built environment are of great interest to public health researchers as they are deemed to be modifiable through intervention and policy. Therefore, the produced results have the potential to directly inform the potential effectiveness and viability of interventions designed by public health researchers.
Overall, this project will be beneficial to the wider public. The health consequences of child and adult obesity remain one of the biggest health burdens today, associated with financial and emotional cost to the state and families. Any effort to understand and prevent the development of childhood obesity, avoiding the long-term health consequences will be beneficial to the society as a whole.