Genomic epidemiology of infectious disease outbreaks

Advances and the growing accessibility of whole genome sequencing of pathogens presents an opportunity for infectious disease epidemiology to expand the toolset currently used to combat infectious disease outbreaks. As the genome contains the complete information that can be obtained about the relationships between isolates, methods developed around the genome can provide us with insight into the history of an outbreak, help us analyse ongoing outbreaks, and enable the detection of hidden outbreaks, associated for example with a particular strain of a pathogen gaining resistance to an antibiotic. This PhD will aim to address three main themes.
Detecting outbreaks, and rapidly quantifying the effect of a mitigation strategy on a pathogen population is an important problem in infectious disease epidemiology. The first theme will focus on local phylodynamics. Phylodynamic methods uses the coalescent process and other stochastic processes to infer the population size history of a pathogen from a sample of genomic data. However, almost all existing phylodynamic methods assume that the whole population follows the same dynamical history. This theme will focus on developing models and statistical methodologies to infer and detect changes within separate subsets of a pathogen population.
From a statistical point of view, existing Bayesian phylodynamic methods are often inefficient, and the priors not elicited in any way that would take a mechanistic model into consideration. The second theme will focus on investigating statistically efficient Bayesian phylodynamic methodologies, as well as how epidemiologic models can help provide better and more principled priors for phylodynamics. New methods of inference (e.g.: particle MCMC) will be applied to improve upon existing phylodynamic tools.
In infectious disease epidemiology, it is often important to consider spatial as well as temporal information. Genetic data itself does not inherently encode any such information itself but is directly affected by it. The third theme will involve integrating genetic sequencing data of pathogens with spatial statistics to enable a better reconstructing of the geographical routes by which a pathogen spread between locations.
The context of the research - Genomic data is increasingly available and has strong potential to complement epidemiological data to help us understand and control infectious disease. However, this potential is currently unrealised due to a lack of methodology that integrate genomic data into an epidemiological framework.
The aims and objectives of the research - The aims of this project are to develop new methods of analysis for genomic data of infectious diseases. This includes the inference of past population sizes, the detection of lineages with different phylodynamic properties, and the reconstruction of geographical routes of spread.
The novelty of the research methodology - The project is based on novel phylodynamic models, and makes use of the latest methods for the inference of parameters under these models.
The potential impact, applications, and benefits - The methods will be applied to several datasets in collaboration with our external partner, PHE. The methods will be implemented and released as open source software which will be useful for the increasingly large number of scientists working in the field of genomic epidemiology.
How the research relates to the remit - The project is highly interdisciplinary, making use of the latest mathematical, statistical and computational methods to reveal insights in infectious disease epidemiology and public health.
Research areas; Healthcare technologies, Mathematical Sciences
External Partner - PHE/NIHP

Impact from the MathSys CDT will arise from three separate mechanisms, each of which will generate a spectrum of academic, economic and societal impacts.

1) Most prominently, this CDT will create the next generation of quantitative researchers that are trained in the necessary skills and techniques to make substantial impact in academia, industry and government agencies. Creation of skilled researchers with a broad scientific outlook will have a number of beneficiaries. We expect that our students will be in high demand within academia and will be the researcher leaders of tomorrow. In addition, many of our brightest students post-PhD are now moving out of academia to research positions within industry or government agencies; such students are likely to generate substantial financial impact within industry and societal benefits within government agencies. By encouraging strong collaboration with our external partner organisations throughout their training, our PhD students will have a broad insight into the impact that mathematics can bring, and the routes through which academic excellence can be translated into meaningful applied outputs with impact. The assembled team of supervisors has an excellent track-record of supporting and training high calibre PhD students with skills that are in demand both within and outside of academia.

2) More immediate economic and societal benefits will accrue from the direct interaction of our students with external partners that is an integral part of their training. We anticipate that 4-6 students per cohort will undertake a PhD that is co-supervised by one of our external partner organisations; in addition all students during their MSc year will partake in one of several group projects led and supported by one of our external partners. In both cases, research will be focused towards real-world problems that are of current concern to the partners. It is anticipated that through these close interactions our students will develop methodologies and results that will address real-world problems. These new solutions to particular challenging real-world problems from external partners are likely to have substantial industrial, economic or societal benefits as they directly tackle prominent and pressing issues set by those with the greatest knowledge of the real-world challenges. Impact will therefore be generated through direct problem-solving research with a number of the UK's leading organisations.

3) Finally, we envisage that the mathematical techniques that are developed in the context of one real-world problem will have wider benefit to other academic fields. Although the immediate beneficiaries are likely to be other academics who will gain from an increased repertoire of tools and techniques, in the longer term these insights are likely to lead to new applications that feed back into industry, finance and society in general. The transdisciplinary nature of our MathSys CDT will facilitate such interactions, promoting the exchange of ideas between diverse subject areas. We firmly believe that such cross-fertilisation of ideas will be a feature of the MathSys CDT, where students are united by common goals of quantitative understanding and prediction and a common language of mathematics. We therefore expect rapid impact in a variety of applied areas, as novel techniques are introduced.