Drug repurposing with deep learning

The proposed project aims to develop deep learning models that can predict with high accuracy likely new indications for drugs that have already been approved for human use. The data that are important for building such models are drug databanks, drugs’ genetic fingerprints in assayed cell lines, original labels. These data are multi-dimensional as well as significant in size, hence the need for incorporating artificial intelligence that has the capabilities to process them. Identifying new indications for existing drugs can both speed up and cut costs for the R&D process. In addition, in our country, the entire drug discovery pipeline has not been in widespread practice due in part to the costly operation and lack of human resources. Focusing on drug repositioning will allow us to essentially sidestep the bottleneck in terms of the operation as well as resource limitations.