Identifying virus-specific signatures in adaptive immune repertoires using deep learning, data augmentation and weak supervision

点击翻译按钮获取中文摘要

The aim of this PhD is to devise deep learning approaches to predict the target viral antigenof individual immune receptors within the immune repertoire of an individual. Given the lackof labelled training data (i.e. repertoire sequences pre-labelled with their target antigen), thestudent will investigate data augmentation and weakly supervised learning strategies. S/hewill also investigate virus-specific signatures of receptors from a germline gene perspective,starting with gene IGHV1-69, which occurs in many neutralising antibodies against virusessuch as HIV. Computational methods development will be complemented by theexperimental validation of predictions throughout the PhD.