EPPuGNN

Prediction and identification of essential proteins are vital for drug development, disease analysis and to better understand cell biological processes, as these proteins are vital for maintaining cell development and growth. In-lab experiments to identify these proteins are time-consuming and expensive; therefore, some machine learning methods were proposed. We review and analyze several prior methods, point out the shortcomings of these methods and reproduce experiments on S. cerevisiae, H. sapiens, D. melanogaster, and M. musculus using state-of-the-art GNNs. You can find software tools, source code, datasets and other relevant artifacts at the GitHub repository of the project.