The biosocial archive: transforming lifecourse social research through the incorporation of epigenetic measures
Lead Research Organisation:
University of Bristol
Department Name: Social Medicine
Abstract
Identifying and understanding the many influences that act across different stages of the life course on social, behavioural and health outcomes continues to present a challenge. This project aims to apply new technologies in epigenomics to interrogate this issue. We propose that the epigenome (markings added to the DNA sequence that alter how genes are regulated) act not only as an index of previous exposures (such as early life adversity, smoking or stress) - a biosocial archive - but may be intimately linked to, and predictive of, future wellbeing.
This project will explore the relationship between epigenetic marks and a variety of different exposures in early life and across childhood and early adulthood. These epigenetic signatures will then be related to social, behavioural and health outcomes. The study will be based upon the well-established Avon Longitudinal Study of Parents and Children (ALSPAC).
Understanding the role of the epigenome and its place in the life course may help to identify when best to intervene to avoid adverse consequences of early life events on long term outcomes. This will have an important impact in informing future policy.
This project will explore the relationship between epigenetic marks and a variety of different exposures in early life and across childhood and early adulthood. These epigenetic signatures will then be related to social, behavioural and health outcomes. The study will be based upon the well-established Avon Longitudinal Study of Parents and Children (ALSPAC).
Understanding the role of the epigenome and its place in the life course may help to identify when best to intervene to avoid adverse consequences of early life events on long term outcomes. This will have an important impact in informing future policy.
Planned Impact
The biosocial archive project develops a new paradigm - that the epigenome can index a wide range of social, environmental and behavioural exposures, that in turn may influence social, behavioural and health outcomes across the life course. The proposed project is innovative and has the potential to have high scientific impact. The overarching aim of this project is to define which stages of the life course are most sensitive to specific biologically effective exposures which will inform the policy making evidence base for future interventions.
Researchers from many disciplines will benefit. The work will contribute to our international profile in innovative epidemiology and public health research, build links between academic disciplines and enhance the knowledge economy. In addition it will provide excellent interdisciplinary training opportunities and improve teaching and learning within our institution. The wider impact will also be reflected in enhanced opportunities for public engagement in research-related societal issues.
The project directly addresses the ESRC priority area "influencing behaviour and informing interventions", through allowing identification of potentially causal links between modifiable influences at particular stages of the life course, which could be the target of interventions, and a wide range of social outcomes. As much of this research will address issues of social inequality, this will also feed into the priority area "a vibrant and fair society".
Researchers from many disciplines will benefit. The work will contribute to our international profile in innovative epidemiology and public health research, build links between academic disciplines and enhance the knowledge economy. In addition it will provide excellent interdisciplinary training opportunities and improve teaching and learning within our institution. The wider impact will also be reflected in enhanced opportunities for public engagement in research-related societal issues.
The project directly addresses the ESRC priority area "influencing behaviour and informing interventions", through allowing identification of potentially causal links between modifiable influences at particular stages of the life course, which could be the target of interventions, and a wide range of social outcomes. As much of this research will address issues of social inequality, this will also feed into the priority area "a vibrant and fair society".
Organisations
Publications
Relton CL
(2015)
From stem cells to the law courts: DNA methylation, the forensic epigenome and the possibility of a biosocial archive.
in International journal of epidemiology
Relton CL
(2015)
Data Resource Profile: Accessible Resource for Integrated Epigenomic Studies (ARIES).
in International journal of epidemiology
Richmond RC
(2016)
DNA Methylation and BMI: Investigating Identified Methylation Sites at HIF3A in a Causal Framework.
in Diabetes
Richmond RC
(2016)
Challenges and novel approaches for investigating molecular mediation.
in Human molecular genetics
Richmond RC
(2015)
Prenatal exposure to maternal smoking and offspring DNA methylation across the lifecourse: findings from the Avon Longitudinal Study of Parents and Children (ALSPAC).
in Human molecular genetics
Richmond RC
(2014)
Approaches for drawing causal inferences from epidemiological birth cohorts: a review.
in Early human development
Sharp GC
(2015)
Maternal pre-pregnancy BMI and gestational weight gain, offspring DNA methylation and later offspring adiposity: findings from the Avon Longitudinal Study of Parents and Children.
in International journal of epidemiology
Simpkin AJ
(2015)
Longitudinal analysis of DNA methylation associated with birth weight and gestational age.
in Human molecular genetics
Simpkin AJ
(2017)
The epigenetic clock and physical development during childhood and adolescence: longitudinal analysis from a UK birth cohort.
in International journal of epidemiology
| Description | Objective 1. To define epigenetic (DNA methylation) signatures that index specific social, behavioural and biological exposures. We have investigated epigenetic signatures for a broad range of exposures predominantly in cord blood but also some across the life-course at ages 7 and 15-17. Measures included maternal BMI, maternal weight gain during pregnancy, maternal depression, domestic violence, alcohol use, previous miscarriages, maternal blood lead and cadmium levels, paracetamol use, various indicators of socio-economic position, overall maternal diet, home air quality, maternal blood vitamin D levels and physical activity. For all but maternal BMI, site specific associations with cord blood DNA methylation were weak. We investigated several methods for testing associations with DNA methylation data aggregated across genomic regions. These methods appeared to improve power to observe associations in most but not all exposures although associations were still relatively weak. A manuscript describing these results is in preparation. Extremely strong and replicable results were observed in autosomal methylation differences between males and females. Evidence suggests that most sex-specific methylation is due to testosterone exposure during early fetal development. Two manuscripts describing these results are in preparation. Objective 2: To assess the persistence of these signatures across the life course To date we have published three papers which explore the patterns of DNA methylation over childhood and adolescence. These involve methylation sites which are related to (i) birth weight and gestational age (ii) maternal smoking and (iii) maternal BMI. These papers have contrasting findings. DNA methylation differences at birth which are related to maternal smoking appear to persist over childhood and adolescence. Conversely, differences at birth which are related to birth weight or gestational age appear to resolve during early childhood. Objective 3: Developing methods to analyse changes in epigenetic patterns over time The ARIES project released the first cohort dataset containing three repeated measurements of DNA methylation. Analysis of these data presents methodological challenges, and we have developed models and methods to combat these. In particular we established non-linear multilevel models as a way to resolve two problems: (i) methylation data over time do not tend to be linear over time and (ii) having repeated measurements on the same individual means that the data are not independent of one another and this complexity has to be dealt with. Three papers using these models have been published or are in press. Objective 4: To deliver a workshop to disseminate findings and engage social scientists in this area of research) We have delivered presentations and workshops nationally and internationally to disseminate key findings from the grant, including at the World Congress of Epidemiology in Anchorage, Alaska (August 2014); the Epigenomics of Common Diseases conference in Cambridge, UK (October 2014); a short course for social scientists and epidemiologists at the University of Bristol (July 2014); a two day epigenetics workshop in Pelotas, Brazil to a group of researchers and lecturers from South America (March 2015); a one-day workshop in Bristol to facilitate social scientists and epidemiologists working in this area. |
| Exploitation Route | 1. Further development of methods Further implementation of the developed models for investigating epigenetic patterns both over time and across the genome will continue to have an impact on researchers in epigenetics. An MRC methodology grant has been awarded to researchers in our unit to further develop these methods. The outputs of this will become increasingly valuable as more DNA methylation data becomes available. 2. Statistical modelling of DNA methylation Most analyses of human DNA methylation currently follow the lead of previous epidemiological research for selecting covariates to include in statistical models. Our reports about methylation associations with a broad panel of exposures suggest that most covariates being included in models unnecessarily since most exposures tend to have quite weak effects on DNA methylation. 3. Outcome prediction Research has shown that human populations do not respond homogeneously to many exposures. This is likely to be true for exposure-induced changes in DNA methylation and may be the reason why associations between exposures and DNA methylation across entire populations tend to be quite weak. If true, then it may be possible to combine DNA methylation profiles with exposure data in order to obtain models predictive of later outcomes. For example, we have begun investigating models that combine prenatal domestic violence exposure and cord blood DNA methylation in order to improve prediction of subsequent behavioural phenotypes. 4. Dissemination of findings Delivering national and international workshops has raised the profile of epigenetic research and will benefit the field by increasing interest. |
| Sectors | Healthcare |
| URL | http://europepmc.org/abstract/MED/25869828 |
| Description | The academic outputs (see Key Findings) of this grant played a key role in the award of two further grants to MRC Integrative Epidemiology Unit (IEU) researchers, to investigate further the relationships between epigenetics, early life social and environmental exposures and later health outcomes. The first is a large inter-disciplinary grant led by Caroline Relton to draw together skills of social and biological scientists (http://gtr.rcuk.ac.uk/projects?ref=ES/N000498/1). The second provides funds for a research network of 6 UK Universities, led by Laura Howe (INTERpreting epigenetic signatures in STudies of Early Life Adversity), focusing on early life adversity (http://gtr.rcuk.ac.uk/projects?ref=ES/N000382/1). In the process of analyzing epigenetic datasets, we encountered several technical challenges. Epigenetic signal in these datasets is always accompanied by technical artefacts, and these artefacts must be removed in order to uncover underlying biological relationships. Datasets are becoming increasingly large as coverage of the human genome increases, rendering many analyses impractical given the software and hardware that is typically available. The high cost of data generation necessitates data sharing whilst respecting limitations imposed by ethical and legal obligations. To address these challenges, we have created 'meffil', a software package for the R statistical computing environment implementing memory and time-efficient algorithms for the analysis of large epigenetic datasets. The package is publically available and routinely used by outside research groups, particularly by members of the Genetics of DNA Methylation Consortium (http://www.godmc.org.uk/). Manuscripts describing the package are in preparation. Our published findings using epigenetic data generated using the Infinium HumanMethylation450 BeadChip have contributed to its widespread use in cohort studies. Prior to its availability, epigenetic data was generated by a wide range of complementary but difficult to compare approaches. Directly as a result of the Illumina BeadChip, a number of consortia have formed to obtain sufficient statistical power through data sharing. IEU members are well represented in these consortia, often in leadership roles (e.g. Caroline Relton leads the Genetics of DNA Methylation Consortium). The popularity of BeadChip prompted Illumina to release an updated version, effectively doubling genomic coverage to 3% of genomic CpG sites. |
| First Year Of Impact | 2015 |
| Sector | Other |
| Impact Types | Economic |
| Title | Methodological developments for incorporation of epigenetic measures in social research |
| Description | |
| Type Of Material | Database/Collection of data |
| Year Produced | 2015 |
| Provided To Others? | Yes |