27th April 2021
Blog: The Bloomsbury SET final Symposium explores the importance of Data Science for combatting infectious diseases and AMR
This blog reports on the recent Big Data symposium that aimed to showcase the results of Data Science projects selected during a two-phase sandpit in 2019.
After a brief introduction to The Bloomsbury SET Knowledge Exchange programme by Dr Emma Tomlinson (RVC), Dr Gwen Knight (LSHTM) chaired the symposium, starting with a presentation from our invited guest Dr Laura Merson (IDDO).
Laura provided a topical overview of the importance of data and data access to accelerate responses to global health crises. She exemplified this with the benefits of publishing virus genetic sequences which enabled the development of COVID-19 vaccines and tracking systems, contrasting with clinical trials data, critical for the understanding of a new disease, which does not follow same trend. There are several areas that needs to be better understood for effective public health responses, including getting clinical data from a wide variety of people and places. Laura made the case for better and more timely responses to outbreaks to change the course of epidemics by having a research framework, protocols, adapting tools and building on existing learning and global networks. This can be possible by having the infrastructure and governance to be able to share data collectively.
Laura’s talk was followed by four presentations of projects funded by the Bloomsbury SET:
Dr Michael Delves from LSHTM presented results on using AI machine learning image analysis routines to analyse drug treated Malaria parasites to accelerate anti-malaria drug discovery. Using knowledge that transmission-blocking drugs affect the morphology of the parasite cells, Michael was able to test off-the-shelf machine learning solutions to analyse and more importantly, cluster microscopic images of cells treated with different drugs. This work is helping to prioritise which molecules have a better chance to be developed into drugs that inhibit transmission.
Professor David Brodbelt from RVC is working on linking data for companion animals (dogs and cats) from veterinary practices. This connects electronic patient records from VetCompass, and culture and sensitivity data from private laboratories. The overall aim is to use Machine Learning in a consistent framework to monitor emergency of AMR in relation to overall prescribing levels of drugs. David explored the challenges related to data integration due to ways the clinical information is collated and sees great potential to add information from other types of data.
Dr Jody Phelan from LSHTM spoke about the development of a framework to collect and curate genomic data. The current project is focused on Tuberculosis, but methodology can be expanded to any other organism. He presented the webserver TB-Profiler, and explained how they are using machine learning on Whole Genomic Sequencing and laboratory-based drug susceptibility testing data to build a pipeline, and framework to collate the data that can inform resistance mutations arising from use of newer drugs. His crowdsourcing approach relies on users to provide data, and this is a vital part to find mutations in a real-time manner rather than collecting it over time. This work can support clinical decision making as new drugs are used.
Dr Martin Walker from RVC works in collaboration with IDDO in neglected tropical diseases caused by soil-transmitted helminths. Efforts to scale-up control of these diseases are based on strategies such as mass-drug administration, but little is known of the efficacy of these programmes. There are also concerns regarding the development of AMR. The project is building a data-share platform on schistosomiasis to facilitate research on treatment responses from patients. The team is developing tools to enhance the findability of data, and new statistical methods to improve understanding of the distribution of individual responses in different communities, rather than the average.
A final panel discussion, excellently chaired by Gwen, dealt with challenges such as:
- the contribution from research and clinical communities to data sharing
- curation, standardisation, and harmonization of data sets
- linkage to interdisciplinary data
- the value of open access
- gaps in data sets, with gender and economic inequalities
The Bloomsbury SET is pleased to see the progress on these projects to bring innovation on the ways we acquire, use, share and value these resources, that are vital for the understanding and control of infectious diseases and AMR.
The recording of this event is available and you can watch here.
Image: Screenshot of panel discussion with Dr Gwen Knight, Dr Martin Walker, Dr Laura Merson (top) and Dr Michael Delves, Professor David Brodbelt and Dr Jody Phelan (bottom).
Dr Adélia de Paula, Knowledge Exchange Officer