Bayesian Collective Markov Random Fields for Subcellular Localization Prediction of Human Proteins

摘要: Advanced biotechnology makes it possible to access a multitude of heterogeneous proteomic, interactomic, genomic, and functional annotation data. One challenge in computational biology is integrate these data enable automated prediction the Subcellular Localizations (SCL) human proteins. For proteins that have multiple biological roles, their correct silico assignment different SCL can be considered as an imbalanced multi-label classification problem. In this study, we developed Bayesian Collective Markov Random Fields (BCMRFs) model for multi-SCL Given set unknown corresponding protein-protein interaction (PPI) network, SCLs each protein inferred by its interacting partners. To do so, PPIs, adjacency features, perform transductive learning on re-balanced dataset. Our experimental results show spatial improves prediction, especially with few annotated instances. approach outperforms state-of-art PPI-based feature-based method