Abstract

Infectious disease differs from other human diseases in its potential for unpredictable and explosive spread. With a changing climate and increasing globalisation, the recognition and prevention of infectious outbreaks is becoming a major concern for health authorities. Antibiotic resistant infections are an increasing concern, with the WHO estimating that the rate of fatal multi-drug resistant bacterial infections could increase to 10 million by 2050, and one of the first cases of a death due to a bacterial infection that was resistant to all available antibiotics occurring early last year. The prevention of outbreaks of infectious disease relies on early warning, as failure to detect an emerging superbug before it has become widely dispersed makes effective elimination by clinical intervention more difficult.

As the cost of genome sequencing decreases, and sequencing technologies become more portable, the option of monitoring infectious diseases using DNA sequencing becomes more realistic. This improvement in technologies has allowed the development of a large number of surveillance programs which collect bacteria from patients in clinics around the world and sequence their genomes. These efforts allow us to understand which bacteria commonly infect us, how often infection results in disease, and whether there are any features in the DNA of these bacteria that allow us to predict which bacteria will be especially problematic for the patient.

Dr Nicole Wheeler will talk about her work using machine learning to facilitate and improve this approach. She will present an algorithm she has developed to detect Salmonella that are more likely to cause bloodstream infections rather than food poisoning, illustrating how the model works and what this teaches us about this disease. She will also outline a new project which aims to detect antibiotic resistant superbugs in the clinic, focussing on the methodological advances that need to be made in the field in order to develop a reliable method that will generalise well into the future.

Speaker: Dr Nicole Wheeler is a postdoctoral fellow at the Sanger Institute. She specialises in using bioinformatics and machine learning to identify patterns of patterns of mutations in bacterial DNA associated with the ability to cause severe disease.