Modelling the genetics of time-varying phenotypes using dynamic Bayesian networks
This is an exciting opportunity to join a multidisciplinary team to develop dynamic Bayesian networks of BMI trajectories leading to childhood obesity. The project builds on cutting edge research developed by the University of Surrey modelling BMI dynamics in childhood and Bayesian semi-mechanistic modelling of dynamic processes.
Start date1 July 2021
The funding package for this studentship award is as follows:
- Full UK/EU tuition fee covered
- Stipend at £15,285 p.a.
- Conference and training stipend of £1,000 p.a.
- Project Desktop or Laptop Computer.
Funding sourceThe University of Surrey, Project-led Studentship Award
Childhood obesity, a risk factor for adult cardio-metabolic disease, is defined as a body mass index (BMI) above the 95th percentile. Currently, over 20% of children are obese, and the prevalence is rising. Despite the relatively large heritability (40-86%), the genes controlling the increase and decrease of BMI at different periods of infancy and childhood remain largely unknown.
Recently, we showed that distinct genetic factors control infant and child BMI . This finding led us to speculate that genetic effects are dynamic, varying throughout life. We, therefore, aim to investigate whether BMI trajectories associated with normal, overweight and obese children are controlled by different genes and whether these genes have varying genetic effects.
Despite the significant progress of growth modelling, BMI remains a challenging phenotype because the BMI trajectory is highly nonlinear and controlled by multiple environmental and genetic factors. Current approaches have focused on data-driven models, while semi-mechanistic models  and the incorporation of domain-specific knowledge have been left unexplored.
The focus of this project is to evaluate and develop probabilistic models of BMI trajectories, and apply them to real data to identify genes controlling BMI trajectories. We intend to develop statistical software for the application of these models in the Comprehensive R Archive Network (CRAN) and make it publicly available.
The candidate will join a supervisory team composed of:
- Dr Alex Couto Alves from the University of Surrey,
- Dr Naratip Santitissadeekorn from the University of Surrey
- Prof. John Holloway from Southampton University, a Genetic Epidemiologist with extensive experience leading large studies and modelling longitudinal phenotypes.
In collaboration with:
- Prof. John Wright, Epidemiologist and Director of the Born in Bradford (BiB) Study;
- Ms Gillian Santorelli, Principal Statistician at BiB and expert modelling longitudinal data at BiB.
Related linksDaily Mail obesity article Applied and Computational Measurable Dynamics book
MSc or equivalent in statistics, bioinformatics, data science, computer science, genetic epidemiology. Experience with either Bayesian data analysis, longitudinal modelling, genetic data analysis, statistical or machine learning modelling is required.
A solid background in any of the programming languages R, Python or C is required.
This studentship is only for UK/EU applicants.
TOEFL or IELTS test is required.
The programme will be the Biosciences and Medicine PhD.
How to apply
Applications can be made through the Biosciences and Medicine PhD programme page at the University of Surrey.
1. Couto Alves, A. et al (2019) GWAS on longitudinal growth traits reveals different genetic factors influencing infant, child, and adult BMI. Science Advances. 5, eaaw3095.
2. Santitissadeekorn, N. et al (2020) Approximate filtering of conditional intensity process for Poisson count data Comput. Stat. Data Anal. 144, 106850
3. Bolt & Santitissadeekorn (2013) Applied and Computational Measurable Dynamics, SIAM
4. Arshad, S., Hodgekiss, C., Holloway, J.W., et. al. (2018) Lung Function Trajectories From Childhood To Young Adulthood. JACI
5. Fairley, L., (…), Wright, J. (2013) Describing differences in weight and length growth trajectories between white and Pakistani infants in the UK. Arch Dis Child.
6. Arshad, S. et. al. (2020) Cohort Profile: The Isle of Wight Whole Population Birth Cohort. Int J Epidemiol.
7. Wrigth, J. et. al. (2013) Cohort Profile: The Born in Bradford multi-ethnic family cohort study. Int J Epidemiol.
Our research focuses on the genetic determinants of adult health and disease. We collaborate in large international consortia running large scale genetic studies of human disease and phenotypes. We apply and develop computational and statistical methods for understanding the aetiology of adult health and disease. We are part of the Centre for Mathematical and Computational Biology where our research on semi-mechanistic models of dynamic processes is based.