Causal associations of body composition and disease-specific mortality in contrasting populations
Obesity is a major cause of premature disease and mortality but different components of body composition have different associations with disease risk. Moreover, most previous evidence is from studies of high-income populations of European ancestry. This project aims to systematically assess the causal relevance of different markers of body composition to multiple diseases and causes of death in the Mexico City Prospective Study (MCPS), and to compare with estimates seen in the UK Biobank (UKB).
The MCPS includes 150,000 Mexican adults followed for two decades with detailed baseline information on socio-demographic factors, lifestyle characteristics, and physical and biological measurements (including NMR metabolomic, genetic array and exome data in the whole cohort). It is also anticipated that some proteomic data on MCPS participants will become available during the course of this project. The UKB includes 500,000 adults with detailed phenotypic, biological and genetic data in a UK population with lower levels of adiposity who have been followed for about 15 years.
The specific aims of this project will be subject to student interest and discussion with the supervisors but could involve:
- Mendelian randomisation approaches (one- and two-sample methods) to assess causality of associations between body composition traits and specific non-fatal or fatal diseases in these contrasting populations.
- Mediation analyses to investigate how different genetically-determined markers of body composition and circulating biomarkers mediate the relationships with risks of specific diseases and mortality.
- Compare associations by ancestry and other characteristics.
RESEARCH EXPERIENCE, RESEARCH METHODS AND TRAINING
The student will gain experience in non-communicable diseases epidemiological research, genetic epidemiology and analysis of large-scale prospective data. They will develop skills in conducting systematic literature reviews, study design for causal inference in a general population context, statistical programming and data analysis, including different types of mediation analyses, and presentation skills. The student will be supported to publish peer-reviewed papers emerging from their DPhil.
FIELD WORK, SECONDMENTS, INDUSTRY PLACEMENTS AND TRAINING
Training in advanced statistics, epidemiological methods, programming, and scientific writing will be provided. Attendance at seminars, workshops and courses provided by the Department and University will also be encouraged. There will be opportunity to present research work at relevant international/national conferences.
The ideal candidate will have a Maste'rs degree in a relevant area (e.g. statistics/genetic epidemiology/biomedical or life sciences) and proficiency with programing analyses in STATA, R or SAS packages.