Xiangxiang Zeng1, Xuan Zhang2, Yuanlu Liao3, Linqiang Pan4. 1. Department of Computer Science, Xiamen University, Xiamen 361005, China. Electronic address: xzeng@xmu.edu.cn. 2. Department of Computer Science, Xiamen University, Xiamen 361005, China. Electronic address: xuanzhang@foxmail.com. 3. Department of Computer Science, Xiamen University, Xiamen 361005, China. Electronic address: yuanluliao@stu.xmu.edu.cn. 4. Key Laboratory of Image Information Processing and Intelligent Control, School of Automation, Huazhong University of Science and Technology, Wuhan 430074, China. Electronic address: lqpan@mail.hust.edu.cn.
Abstract
BACKGROUND: Deciphering the genetic basis of human diseases is an important goal in biomedical research. There is increasing evidence suggesting that microRNAs play critical roles in many key biological processes. So the identification of microRNAs associated with disease is very important for understanding the pathogenesis of diseases. METHODS: Two multipath methods are introduced to predict the associations between microRNAs and diseases based on microRNA-disease heterogeneous network. The first method, HeteSim_MultiPath (HSMP), uses the HeteSim measure to calculate the similarity between objects and combines the HeteSim scores of different paths with a constant that dampens the contributions of longer paths. The second one, HeteSim_SVM (HSSVM), uses the HeteSim measure and the machine learning method used to combine HeteSim scores instead of a constant. RESULTS: We use the leave-one-out cross-validation to evaluate our novel methods, and find that our methods are better than other methods. We achieve an area under the ROC curve of 0.981 and 0.984 respectively. We also check the top-10 most similarity of microRNAs-diseases associations and find that our predictions are reasonable and credible. CONCLUSIONS: The encouraging results suggest that multipath methods can provide help in identifying novel microRNA-disease associations, and guide biological experiments for scientific research. This article is part of a Special Issue entitled "System Genetics". Guest Editor: Dr. Yudong Cai and Dr. Tao Huang.
BACKGROUND: Deciphering the genetic basis of human diseases is an important goal in biomedical research. There is increasing evidence suggesting that microRNAs play critical roles in many key biological processes. So the identification of microRNAs associated with disease is very important for understanding the pathogenesis of diseases. METHODS: Two multipath methods are introduced to predict the associations between microRNAs and diseases based on microRNA-disease heterogeneous network. The first method, HeteSim_MultiPath (HSMP), uses the HeteSim measure to calculate the similarity between objects and combines the HeteSim scores of different paths with a constant that dampens the contributions of longer paths. The second one, HeteSim_SVM (HSSVM), uses the HeteSim measure and the machine learning method used to combine HeteSim scores instead of a constant. RESULTS: We use the leave-one-out cross-validation to evaluate our novel methods, and find that our methods are better than other methods. We achieve an area under the ROC curve of 0.981 and 0.984 respectively. We also check the top-10 most similarity of microRNAs-diseases associations and find that our predictions are reasonable and credible. CONCLUSIONS: The encouraging results suggest that multipath methods can provide help in identifying novel microRNA-disease associations, and guide biological experiments for scientific research. This article is part of a Special Issue entitled "System Genetics". Guest Editor: Dr. Yudong Cai and Dr. Tao Huang.