Leveraging molecular-QTL co-association to predict novel disease-associated genetic loci using a graph convolutional neural network

This project aimed to identify novel disease-associated loci by leveraging the co-association between genetic loci and molecular traits (DNA CpG-site methylation status and RNA expression). Using a graph convolutional neural network, we encoded this co-association in the form of embeddings -- which we provide here. These embeddings can be used to recommend novel disease-associated loci, based on those already known to be associated with a particular disease. Please refer to README.txt for further information.