Early career fellowship 2022
The MRC Integrative Epidemiology Unit (MRC IEU) is seeking to support talented and motivated researchers who wish to consolidate their postgraduate research through a short-term training position, to be held in the MRC IEU.
The IEU Early Career Research fellowships offer a salaried position on a fixed term contract of up to 12 months (ending on or before 31 March 2023). This is a training position with mentoring from senior academic staff. The purpose is for the candidate to develop new skills or develop an individual research vision through doing pilot research projects.
Applicants will have recently completed (or may be working towards completion at the time of application) a PhD or MSc in a discipline relevant to causal epidemiology. Examples of relevant disciplines include mathematics, statistics, computer science, causal or translational epidemiology (including veterinary epidemiology), data mining and bioinformatics, psychology, genetics, biomedical science, economics, medicine or policy studies. Applicants will be strongly motivated to develop their skills and knowledge in a topic aligned to one of the IEU research programmes (see list below).
We are looking for applications that clearly describe the following:
- Benefit to be gained by the applicant through this training position
- Alignment with the mission of the IEU
- Quality and relevance of the proposed activity
- Evidence of appropriate track record for the proposed activity
The deadline for applications is 30th March 2022. To apply visit the Working at Bristol page.
Update (18 March):
New cancer epidemiology topics added. find these under the epigenetic epidemiology topic.
Mendelian randomization
Key team members: George Davey Smith, Gibran Hemani
Topic description: We are keen to hear from applicants who would like to get involved in research relating to any of the following 3 areas;
- Apply Mendelian randomization (MR) methods to studies of disease progression, addressing hypotheses of direct therapeutic relevance
- Combine MR approaches with other methods aimed at strengthening causal inference, contributing to the development of “triangulation of evidence” methods
- Investigate problems introduced into MR by the expansion of the size of GWAS and consequent incorporation of variants influencing upstream as instruments for downstream exposures, and consider ways to alleviate these problems
Please get in touch if you'd like to discuss your project ideas or hear more about suggestions we have. We welcome projects that may cut across more than one of the MRC IEU programmes.
Please contact Prof George Davey Smith or Gibran Hemani in the first instance (KZ.Davey-Smith@bristol.ac.uk) (g.hemani@bristol.ac.uk)
Statistical methods for improving causal analyses
Key team members: Kate Tilling, Rachael Hughes, Jon Heron
Topic description: We are keen to hear from applicants who would like to get involved in research relating to any of the following 3 areas;
- Develop and test ways to deal with missing data in MR using inverse probability weighting or multiple imputation
- Develop and apply methods to jointly model an exposure and an outcome trajectory
- Investigate methods for causal analyses in prognostic studies
Please contact Prof Kate Tilling in the first instance (kate.tilling@bristol.ac.uk)
Data mining epidemiological relationships
Key team members: Tom Gaunt
Topic description: We are offering the opportunity to develop skills in drug discovery, research software engineering and/or novel research data platforms. You are welcome to develop your own project ideas in this area. The following examples illustrate some of the opportunities available:
- From target genes to target pathways: developing a pathway-MR approach to identify novel drug targets. Mendelian randomization is widely used to predict the effect of drug targets on diseases by using genetic variants that influence those targets. However, the data available to carry out these analyses do not provide complete coverage of all genes/proteins that could be targeted by drugs. One potential approach to addressing this is to consider pathways as the unit of investigation instead of the individual proteins themselves. Such an approach could highlight the pathways that merit further investigation and enable identification of alternative targets if a primary result is not a tractable drug target. This project would focus on developing MR methods that integrate proteins within a pathway/network.
- “Tera-scale” OpenGWAS: Novel database architectures for performant cloud-based genetic data sharing. The OpenGWAS platform contains over 200 billion genetic association results and is continually growing. New datasets emerging over the next two years are predicted to push this beyond 1 trillion. The database currently uses an ElasticSearch database hosted in Oracle Cloud for rapid queries via a Python API and R packages. This is complemented by a static repository of VCF-format files for download and local HPC use. This project would focus on evaluating and implementing novel database technologies and architectures to support performant cloud-based data access. A core objective would be to support data federation so that we can enable integration with other databases.
- OpenGWAS and EpiGraphDB Python ecosystems: Novel Python packages to support genetic epidemiology and drug discovery. The OpenGWAS and EpiGraphDB platforms are complemented by an array of R packages to support genetic epidemiological analyses, knowledge discovery and data integration. These tools are already used by a wide array of researchers around the world. However, researchers in some domains primarily use Python in their workflows. This project would involve the development of new Python packages, both to complement existing R packages and to provide new functionality.
Please get in touch if you'd like to discuss your project ideas or hear more about suggestions we have. We welcome projects that may cut across more than one of the MRC IEU programmes.
Please contact Tom Gaunt in the first instance (tom.gaunt@bristol.ac.uk)
Epigenetic epidemiology
Key team members: Caroline Relton, Rebecca Richmond, Richard Martin
Topic description: We are keen to hear from applicants who would like to get involved in research relating to any of the following areas;
Epigenetic epidemiology:
- Using DNA methylation as an exposure indicator to understand disease risk
- Using DNA methylation in disease prediction or prognosis, where possible through augmenting other data sources that have predictive value
- Using causal inference methods and DNA methylation data to understand pathways to disease
Cancer epiemiomiology in our affiliate Integrative Cancer Epidemiology Programme (ICEP):
- Identify novel, potentially modifiable factors that are causal in cancer risk and prognosis, and identify novel drug targets, through analysis of large-scale, international population studies.
- Clarify molecular mechanisms underlying effects of modifiable exposures on cancer risk and prognosis by integrating population and basic sciences.
- Identify novel biomarkers for cancer early detection and screening, risk prediction and prognosis through multi-‘omic approaches.
- Translate causally relevant exposures and predictive biomarkers into cancer primary and tertiary prevention interventions.
Please get in touch if you'd like to discuss your project ideas or hear more about suggestions we have. We welcome projects that may cut across more than one of the MRC IEU programmes.
Please contact Prof Caroline Relton in the first instance (caroline.relton@bristol.ac.uk) and Richard Martin for interest in cancer epidemiology topics (Richard.Martin@bristol.ac.uk)
Reproductive and cardio-metabolic health
Key team members: Deborah Lawlor, Carolina Borges
We are keen to hear from applicants interested in research in the following areas:
Project 1: The role of maternal circulating proteins in gestational diabetes. The Born in Bradford cohort recruited ~13,500 women during their pregnancy between 2007-2020. Very detailed information has been collected on them and their offspring since that time. This project would use proteomic data (450 quantified serum proteins) on a subset of 1000 women (50% South Asian and 50% White European origin), including all cases of women with gestational diabetes (diagnosed with a gold-standard oral glucose tolerance test). In this pilot study, the main analyses would be to: (a) compare protein levels between South Asian and White European women; (b) explore associations of proteins with gestational diabetes and offspring birth weight. There would be additional opportunities if time allowed, including: (c) comparing associations of genetic variants from genome-wide association studies of proteins in non-pregnant women and men, with the same associations in these pregnant women; (d) undertaking preliminary genetic, Mendelian randomization, analyses.
Project 2: Triangulation to improve causal inference. Triangulation refers to the integration of evidence from different types of data and/or with different study designs or analytical methods, where each data type/methods have different key sources of bias. The idea is that if results from these different approaches point to the same causal answer then it is more likely that is the true causal answer than relying on the result from just one of the approaches. Epidemiological studies are increasingly described as using triangulation, but varying from using two different statistical approaches in one data set to a systematic approach to try and identify all relevant evidence and compare different results across many different data/methods. This project would review triangulation studies published over the last 10 years and describe the approach to triangulation across different publications and how the causal question is defined and risk of bias explored in each publication. There would also be the opportunity to develop analytical methods for combining results from different methods and complete a triangulation study to address a specific causal question.
Please get in touch if you'd like to discuss your project ideas or hear more about suggestions we have. We welcome projects that may cut across more than one of the MRC IEU programmes.
Please contact Prof Deborah Lawlor in the first instance (d.a.lawlor@bristol.ac.uk)
Translation of behavioural interventions in public health
Key team members: Marcus Munafò, Angela Attwood
Topic description: We are keen to hear from applicants interested in research in the following areas:
- Developing and delivering novel interventions to improve public health. The work will build on previous findings from observational and experimental studies that have identified causal targets for intervention. The areas of focus include reducing alcohol consumption in heavy non-dependent drinkers, informing NHS smokefree site policy and tackling the increase in use of “puff bars” (single-use disposable e-cigarettes) among adolescents. These projects will be conducted in collaboration with local hospitals and Bristol City Council and provide experience of working at the applied end of the translational pathway. Contact Angela Attwood in the first instance (angela.attwood@bristol.ac.uk)
-
Applying Mendelian randomization (MR) methods to studies of the behavioural determinants of physical and mental health, or influences on health behaviours.
- Applying multivariable MR to dissect the unique contributions of multiple health behaviours, components of behavioural exposures, and/or socioeconomic determinants of physical and mental health.
- Developing individual- and population-level interventions that target pathways identified in previous MR studies, with a view to conducting feasibility and pilot studies.
Please get in touch if you'd like to discuss your project ideas or hear more about suggestions we have. We welcome projects that may cut across more than one of the MRC IEU programmes.
Please contact Prof Marcus Munafò in the first instance (marcus.munafo@bristol.ac.uk).
Translational mental health
Key team members: Golam Khandaker, Dheeraj Rai
Topic description: We are keen to hear from applicants interested in research in the following areas:
- Role of inflammation in depression, schizophrenia, and neurodevelopmental disorders.
- Role of inflammation in psychiatric disorder relevant phenotypes such as structural/functional neuroimaging brain changes, cognition, neurodevelopment.
- Role of inflammation in psychiatric and physical multimorbidity.
Please get in touch if you'd like to discuss your project ideas or hear more about suggestions we have. We welcome projects that may cut across more than one of the MRC IEU programmes.
Please contact Prof Golam Khandaker in the first instance (golam.khandaker@bristol.ac.uk)