Hereditary UE Horizon project

HEREDITARY is an interdisciplinary project aimed at improving healthcare by integrating diverse health data to uncover insights into neurodegenerative and gut-brain disorders like ALS, Parkinson’s, and Alzheimer’s. It uses cutting-edge AI, federated learning, and privacy-preserving frameworks to harmonize clinical, genomic, and environmental data.

I started working in this project few months ago where the idea is to implement and fine-tuned transformer models like DNABERT, Nucleotide Transformer and HyenaDNA to process genomic sequences specific to ALS patients. Focused on discovering genetic patterns to aid in the identification of possible novel therapeutic targets. Collaborated on developing a self-supervised framework for processing unlabeled data, improving model performance on large-scale patient datasets like the AnswerALS dataset.