Zoonotic diseases are a major concern for poor livestock keepers in lower & middle income countries where they are exposed to a huge range of pathogens many of animal origin. However, understanding of what pathogens people are infected with in these settings in sub-Saharan Africa (SSA) is almost non-existent, and consequently it is difficult to develop evidence based interventions and strategies to mitigate the zoonotic risks experienced by people at the livestock/wildlife interface. Non-specific fevers present an increasing challenge as most are still diagnosed as malaria due to lack of differential tools at the patient level. This also has major implications for the development and spread of antimicrobial resistance (AMR) with the blanket use of antibiotics to treat fevers in the absence of any other diagnostic capability, even though a large proportion of these may not even be bacterial in origin - a recent study in children in Tanzania identified viral causes of non-malarial fevers in 70% of patients.
Rapid on-site sequencing would be a major step forward, and small portable sequencing devices exist. These have been successfully used in the Ebola  and Zika outbreaks but critically it was already known which strains to target. For livestock diseases and potential zoonoses a more agnostic approach is needed, and obtaining low error rate sequences would allow transmission networks (who infected whom) to be inferred – these are valuable tools for understanding where to direct resources for the control of infectious diseases .
Transmission networks can be inferred from pathogen sequence data due to the fast evolution and generation times, although their fidelity depends upon the number of (non-error) mutations observable at the appropriate temporal and spatial scales.
In this project you will investigate the use of nanopore sequencers with field samples and compare to next generation sequencing. You will develop and evaluate algorithms to extract the most out of the data – minimising sequencing error and inferring transmission networks (including under missing data scenarios), considering a variety of algorithm types including Bayesian inference, machine learning or deep learning .
Initial pathogen samples will include bovine tuberculosis and foot and mouth disease virus from our collection of outbreaks, but there will be opportunity to collect additional samples and try your protocols and algorithms in the field in later years. In particular this project will build upon existing Roslin-African collaborations including the Bamenda Regional TB reference Laboratory studying bovine and zoonotic tuberculosis and extensive experience working in other parts of Cameroon on FMD, BVD, Q fever and other diseases.
This project combines sequencing technology, field epidemiology and computer science / maths & statistics, and the student will be trained in laboratory, field work, computer programming and data science.
 Real-time, portable genome sequencing for Ebola surveillance, Quick et al, Nature 530, 228–232 (2016) http://www.nature.com/nature/journal/v530/n7589/full/nature16996.html
 Supersize me: how whole-genome sequencing and big data are transforming epidemiology, Kao et al, Trends Microbiol. 2014 May;22(5):282-91 http://www.ncbi.nlm.nih.gov/pubmed/24661923
 Emerging Concepts of Data Integration in Pathogen Phylodynamics, Baele et al, Syst Biol (2016) http://sysbio.oxfordjournals.org/content/early/2016/08/18/sysbio.syw054.full