Machine Learning for Antibodies
This little pig is actually two nanobodies bound to a membrane protein complex - PDB structure 7qn9, for those interested! (Created in PyMOL)
PhD student, AI + Antibodies
University of Oxford
Hi there. I'm a current PhD student within the Oxford Protein Informatics Group (OPIG) at the University of Oxford. My work focuses on applying machine learning techniques to speed up and reduce the cost of therapeutic antibody development (drugs which are used to treat a wide range of diseases including cancer, arthritis, and viral infections such as COVID-19). More broadly, I'm interested in all areas of science and machine learning. My Masters project involved analysing rare particle decays observed at CERN, and I have also created a simple exoplanet detection guide with the University of Hertfordshire. Please checkout a slightly more detailed breakdown of my work and interests below, and do give me a message if you'd like to chat!
My current research involves applying structure-based deep learning methods to predict antibody-antigen binding interfaces. Knowledge of these interfaces can help predict the orientation and strength with which an antibody binds its target (antigen), reducing our reliance on expensive experimental techniques. If you are interested in this or similar work, then please do check out some of OPIG's research by clicking the pig below (any papers I publish will be added here)! In addition, all software we create is provided open-source and can found on GitHub.
This little pig is actually two nanobodies bound to a membrane protein complex - PDB structure 7qn9, for those interested! (Created in PyMOL)
To complement my research and improve my business understanding of the world of biotech, I've recently been involved with Creative Destruction Lab (CDL), a start-up accelerator. I worked with CDL as a student consultant, where I helped build revenue forecasts for Hypervision Surgical, an AI-powered medical imaging venture. Previously, I've also taken part in OxAI Labs, where we tried, unsuccessfully(!), to improve the classification accuracy and speed of particle jet images observed at CERN. As I'm clearly not quite ready to leave my physics background completely behind me, I also regularly tutor first-year undergraduate physicists in preparation for their exams.
Immediately after finishing my Masters I spent eight months tutoring and working in an ice cream shop. After saving enough money, I then happily spent it all by disappearing around Asia and Southern Africa for half a year. On returning, I started work at the University of Hertfordshire as a physics research and outreach assistant. Here, I analysed both transit and radial velocity data in a bid to identify new exoplanets. Building on this experience, I then created a comprehensive guide to enable others - particularly school-age students - to do the same using freely accessible online data.
Finally, I have also spent a couple of years working as a Technical Consultant for BT and as a Supply Chain Analyst at Tesco. Though my current research is clearly quite distant from these experiences, I believe the soft skills I learnt have helped me maintain a healthy work-life balance during the PhD so far!