Jacob Bradley is a doctoral researcher based at the University of Edinburgh, working on interdisciplinary projects between statistical learning and genomics. He graduated from the University of Cambridge in 2019 with degrees in Mathematics and Systems Biology, and enjoys applying modern methods from statistics and machine learning research to impactful problems in biology and medicine. He has particular expertise leveraging data from high-throughput technologies including genomic sequencing.
Immunotherapy, the name given to a new class of drugs aiming to bolster the immune system, is a promising treatment option for many types of cancer. However, it doesn't work well for everyone, and so a great deal of research is aimed at predicting which patients will benefit from receiving immunotherapy. A promising source of data for predicting response to immunotherapy comes from genomic sequencing, where samples from a patient's tumour are analysed to understand where their DNA differs from that of normal cells.
One issue in using sequencing data to inform clinical decisions is a scarcity of suitable data. To be specific, there has been a large amount of sequencing data generated for 'generic' cancer patients, but far less for those who have received immunotherapy treatment for their cancer.
In Jacob's research at the Turing, he works to understand how we can combine these data sources to understand the factors influencing response to immunotherapy. Two problems are key: firstly, how do we separate genomic factors strongly associated with survival from cancer in general from those specifically interacting with immunotherapy treatments? And secondly, how do we condense 'high-dimensional' inputs from genomics sequencing (comprising a very large number of measurements) into a single output predictor of response?