PlantPathoPPI: An Ensemble-based Machine Learning Architecture for Prediction of Protein-Protein Interactions between Plants and Pathogens
The identification of PPIs in the plant-pathogen context is crucial for understanding the intricate molecular mechanisms underlying plant defense and pathogen virulence. Due to the time and effort required for experimental PPI identification, computational approaches are becoming increasingly valuable. In this study, we aimed to develop a robust machine learning-based tool for predicting protein-protein interactions (PPIs) in diverse plant-pathogen systems.The study employed a machine learning ensemble model, combining multiple learning algorithms and diverse sequence encodings, to predict PPIs. The resulting ensemble model, PlantPathoPPI, achieved an improved accuracy of approximately 97% compared to individual models. Comparative assessments with existing tools demonstrated the promising potential of PlantPathoPPI. Furthermore, a web-based prediction server (available at http://login1.cabgrid.res.in:5090/) and Python package (available at https://pypi.org/project/plantpathoppi-ml/) were developed to facilitate accessibility for end users. The web server is hosted at the Advanced Supercomputing Hub of Omics Knowledge in Agriculture (ASHOKA), ICAR-Indian Agricultural Statistics Research Institute, New Delhi.
