Multi-Similarities Integration and Network Projection to Identify Human Microbe-Disease Association

Recent studies have shown that human microbes are closely related to diseases. To study the relationship between microbes and diseases is conducive to an in-depth and comprehensive understanding of pathological mechanisms of diseases, thus providing new means for diagnosis, treatment and prevention of diseases. At present, there are two methods. One is the traditional experimental method, but it is expensive and limited, and the other is the computational method, it is quick and practical. In this paper, we use network projection and fusion of Gaussian kernel, functional and cosine similarity to predict the microbe-disease association. In the cross-validation experiment, our method obtains an AUC value of 0.9832. It shows that our method is an effective prediction tool and can provide reliable disease-related microbial candidates for biological experiments.