The HDR UK-Roche Hackathon, held in April 2024, saw 18 PhD students from eight universities across the UK converge at the Wellcome Collection in London for the second joint collaborative training event between Health Data Research UK and the Data & Analytics unit of Pharmaceutical Research and Development at Roche.
Following on from the Summer School last summer, these events are designed to give early career researchers insight into challenges and technologies, which are of importance to the pharmaceutical industry. This year, participants focused on protein function prediction. Proteins are formed from sequences of molecules called amino acids. These amino acid sequences determine the shape and function of proteins and there has been longstanding interest in developing algorithms to predict how proteins work from their amino acid sequence. This could help us to better understand how genetic mutations give rise to different health conditions or allow new therapies to be developed to treat diseases.
Recent developments in artificial intelligence have enabled dramatic improvements in protein function prediction and participants built upon these to devise their own prediction algorithms in a challenge jointly developed by Roche scientist Joshua Payne and HDR UK PhD student Jude Wells.
“Computational approaches to protein fitness prediction and protein design are rapidly advancing fields with great potential for driving breakthroughs in biology and medicine. We were really impressed with the participants’ ability to develop innovative solutions to this challenging task.”
Jude Wells, University College London, HDR UK Hackathon Lead Participants were also given insight into research undertaken at Roche. Since the event, three participants have also gone on to pursue internships at Roche.
The summer school and hackathon were delivered as part of a UKRI Turing AI Fellowship awarded to hackathon co-lead organiser and HDR UK PhD Programme Director, Professor Christopher Yau, who commented:
“I am really pleased that I have been supported by HDR UK and Roche to use my Turing AI Fellowship to support training and career development of early career researchers from across the country. The success of these two events has enabled us to bring together a diverse group of PhD students to address scientific challenges and to gain insight into Roche and I look forward to developing further partnerships in the future.”
The Capacity Building team at HDR UK are looking for organisations to partner to develop similar training events. If you are interested in discussing opportunities, please contact the team for further information (phd@hdruk.ac.uk).
Profile: Amy Francis, Hackathon Participant and Roche Intern
University: Bristol
Area of study: Integrative Cancer Epidemiology PhD Programme at the University of Bristol, funded by Cancer Research UK
Why did you want to participate in the hackathon?
I’ve participated in previous hackathons and was keen to get involved when I heard about this opportunity on the Turing-Roche Slack workspace. For my PhD, I am developing bioinformatics and machine learning approaches to predict the impact of single nucleotide variants in cancer. The hackathon project on predicting the impacts of mutations in proteins aligned perfectly with my research interests and I was keen to meet people who were also interested in this research. I was confident that I could contribute ideas while challenging myself to apply my knowledge from a different angle.
Additionally, I enjoy working on new challenges and collaborating with individuals from diverse and interdisciplinary backgrounds, which brings unique ideas and perspectives to the table. These events have accelerated my learning, enabled me to approach problems from new directions, and allowed me to network with people outside my immediate area of study. I was also excited to learn more about Roche and HDR at the event.
What did you enjoy most about the hackathon?
My favorite part of the hackathon was the unique contributions each team member brought to the challenge. Whether it was biological knowledge, machine learning expertise, or data visualisation skills, we all learned from each other based on our individual strengths. Additionally, hearing about Roche was very insightful. Of course, our team winning the hackathon was a bonus!
What will you be doing as part of your internship?
I’ll be working in Roche’s Early Research & Development team (pRED, Data & Analytics), collaborating with colleagues based in Zurich and Basel. During the six-month internship, I will focus on using machine learning to predict the impact of combinations of mutations in antibody binding site, with the overarching goal of developing enhanced drugs.
What do you plan to do next?
When I return to my PhD in January, I will have a few months to finalise and submit my thesis. Following this, I plan to further develop my expertise in AI research projects within the biomedical field and build meaningful connections in this area. I am excited about continuing my work in this space and contributing to the next generation of AI research in this dynamic and impactful field!
