A deep ensemble model to predict miRNA-disease association

摘要: Cumulative evidence from biological experiments has confirmed that microRNAs (miRNAs) are related to many types of human diseases through different processes. It is anticipated precise miRNA-disease association prediction could not only help infer potential disease-related miRNA but also boost diagnosis and disease prevention. Considering the limitations previous computational models, a more effective model needs be implemented predict associations. In this work, we first constructed miRNA-miRNA similarity network utilizing functional data heterogeneous Gaussian interaction profile kernel similarities based on assumption similar miRNAs with functions tend associated diseases, vice versa. Then, disease-disease using semantic information data. We proposed deep ensemble called DeepMDA extracts high-level features stacked autoencoders then predicts associations by adopting 3-layer neural network. addition five-fold cross-validation, another cross-validation method evaluate performance model. The results show superior methods high robustness.