Most new human diseases, including Ebola and Zika, are viruses that originate from other animal species, often as a consequence of human encroachment into natural environments combined with greater connectivity of urbanizing societies. The threats to wildlife conservation, human health and global security posed by emerging viruses has led researchers around the world to begin systematically sequencing the vast diversity of viruses that circulate in wild and domestic animals, an effort powered by the increasing speed and affordability of viral genome sequencing. However, while genomic databases grow, our abilities to understand the origins of emerging viruses and to anticipate the risk that any single virus poses to humans prior to its emergence remain in their infancy. Moreover, once new predictive tools are developed, they risk being overlooked if they lack sufficient validation or if pathways to their dissemination to the stakeholders that respond to outbreaks are not identified.

We recently developed a series of machine learning algorithms that predict virus-host associations from evolutionary signals in viral genomes. This is an important advance towards a new generation of digital tools for health interventions that capitalize on the influx of viral genomic data. These algorithms also provide a relevant context in which to understand the barriers to implementing new health tools in the real world.

This 4-year PhD project will enhance and validate machine learning models that infer key aspects of viral ecology from viral genome sequences and will conduct field research in Uganda to identify how digital tools should be applied in the context of health emergencies. The candidate will be supervised by a multi-disciplinary team of researchers (Streicker, Babayan and Bunn) and will collaborate more broadly across the University of Glasgow. The output of the PhD will be a set of predictive, genomic data driven models that provide insights into viral ecology and evolution and are tailored for field deployment as part of surveillance or outbreak response programmes.

Key research questions:

(1) Embracing uncertainty: how can we harness the power of large, but poorly annotated public databases to inform machine learning models that predict the animal origins of emerging viruses?

(2) Proof of concept: do field collected samples from wildlife around the world validate predictions from public databases?

(3) Going public: how to move emerging technology from the academic realm to real-world deployment?

Methodology and training

The highly interdisciplinary and international nature of this project represents exceptional doctoral training for candidates seeking to bridge biological, computational and sociological boundaries in the emerging field of One Health. The candidate will have opportunities to learn cutting edge statistical methods including machine learning and bioinformatics, the newest genomic sequencing technologies (metagenomics) and how to design and implement questionnaires and interviews for application in a developing country. These technologies and skills have broad applicability, giving the scholar transferable skills for a variety of career choices. Predicting viral emergence and understanding the risks posed by newly discovered viruses are extremely hot topics at the interface of ecology, evolutionary biology and public health. As genomic data are now the first information available for these viruses, the output of this PhD (a tool to make inferences from genomes) will have significant impact. The resulting high visibility and skills developed will make the candidate well placed for employment in variety of disciplines including fast-growing fields of artificial intelligence, genomics or global health.

Project Team - The Streicker Lab (Institute of Biodiversity, Animal Health and Comparative Medicine & MRC-University of Glasgow Centre for Virus Research) will provide expertise in viral ecology, evolution and emergence, database management and computational analyses.

The Babayan Lab (Institute of Biodiversity, Animal Health and Comparative Medicine) will provide expertise in the machine learning components of the project.

The Bunn Lab (School of Social and Political Sciences) will provide expertise on the social reception of biomedical innovation, the evaluation of new social and technological processes and social theory (including from science and technology studies).

Additional collaborators will include Richard Orton (bioinformatician), Emma Thomson (clinical virologist), Roman Biek (molecular ecologist) and Ke Yuan (computer scientist).

The scholar will be based at the Gilmorehill (Graham Kerr Building) and Garscube (Henry Wellcome Building) campuses at the University of Glasgow. The field components of the project will take place at the Uganda Virus Research Institute in Entebbe, Uganda.