Multi-source data with Laplacian eigenmaps and denoising autoencoder for predicting microbe-disease association via convolutional neural network

Identifying microbes associated with diseases is important for understanding the pathogenesis of diseases as well as for the diagnosis and treatment of diseases. In this article, we propose a method based on a mul-ti-source association network to predict microbe-disease associations, named MMHN-MDA. First, a heter-ogeneous network of multi-molecule associations was constructed based on associations between microbes, diseases, drugs and metabolites. Second, the graph embedding algorithm Laplacian eigenmaps is applied on the association network to learn the behavior features of microbe nodes and disease nodes. At the same time, the denoising autoencoder(DAE) is used to learn the attribute features of microbe nodes and disease nodes. Finally, attribute features and behavior features are combined to get the final embedding features of microbes and diseases, which are fed into the convolutional neural network(CNN) to predict the microbe-disease as-sociation. Experimental results show that the proposed method is more effective than existing methods. In addition, case studies on bipolar disorder and schizophrenia demonstrated good predictive performance of the MMHN-MDA model, and further the results suggest that gut microbes may influence host gene expres-sion or compounds in the nervous system, such as neurotransmitters or alter blood-brain metabolites of the barrier, which in turn affect people's mood or behavior, thereby inducing psychological diseases.