December 6, 2019, by Charlotte Anscombe
Cholera Antibiotic Resistance in Bangladesh (CARE): big data mining and machine learning to improve diagnostics and treatment selection
A new research project will address the need for rapid diagnosis of cholera by developing tools to help early detection, and provide real-time intervention in outbreaks of this deadly disease.
The project will make use of portable, real-time DNA sequencing (using the Oxford Nanopore MinION), together with cloud-based solutions for large data storage and remote data analysis. Researchers will be able to feedback information in real-time at the local level, and share information at the global level.
The research is being led by Dr Tania Dottorini from the School of Veterinary Medicine and a member of the Future Food Beacon leadership team at the University of Nottingham, together with Dr. Muhamad Maqsud Hossain from the North South University, Bangladesh, and is funded by the Global Challenges Research Fund.
Cholera is an acute diarrheal disease caused by eating or drinking contaminated food and water. In 2017, the WHO reported that around 785 million people lacked a basic water drinking service, putting them at risk of diseases like cholera.
In Bangladesh, 40% of the population is estimated to be at risk, and, alongside India, the country has the largest population at risk from cholera outbreaks. Cholera outbreaks need to be predicted and swiftly dealt with – our current inability to do so contributes to over 100,000 estimated cases each year. The Global Taskforce for Cholera Control wants to eliminate cholera by 2030. In order to reach this ambitious target, new methods of diagnosis, detection and response are needed.
Currently, traditional microbiological testing is resource-intensive, and the referral of cases to laboratories is challenging due to the lack of transport, storage, and testing facilities (especially from rural areas). Researchers involved in CARE Bangladesh will be able to diagnose infection, identify specific cholera strains, detect antibiotic resistance profiles and suggest treatment without the need for intensive microbiological testing by making use of machine learning and statistical modelling accessed via remote computing.
The project brings together local communities, specialised institutes, health educator teams, international development agencies, academics in Bangladesh, USA, and UK, and has the support of the Bangladesh Ministry of Science and Technology and UNICEFF. Using machined learning, in combination with data mining, will allow Bangladeshi health workers to identify patterns for early detection and control of the disease.
Dr Dottorini said: “This is a great opportunity to overcome the barriers that currently reduce the possibility for Bangladeshi people to access diagnostics and treatment solutions which could really make the difference. I am happy and humbled that the Ministry of Science and Technology and UNICEFF stood up to support this initiative and I am really thankful to the GCRF and the University of Nottingham for funding this project. A big challenge, but exciting times ahead.”