MSCNE:Predict miRNA-Disease Associations Using Neural Network Based on Multi-Source Biological Information

The important role of microRNA (miRNA) in human diseases has been confirmed by some studies. However, only using biological experiments has greater blindness, leading to higher experimental costs. In this paper a high-efficiency algorithm based on a variety of biological source information and applying a combination of a convolutional neural network (CNN) feature extractor and an extreme learning machine (ELM) classifier is proposed. Specifically, the semantic similarity of diseases, the gaussian interaction profile kernel similarity of the four biological information of miRNA, disease, long non-coding RNA (lncRNA) and environmental factors (EFs), and the similarities of miRNAs are fused together. Among them, miRNAs similarity is composed of miRNA target information, sequence information, family information, and function information. Then, the dimensionality of the data set is reduced by the autoencoder (AE). Finally, deep features are extracted through CNN, and then the association between miRNA and disease is predicted by ELM. The experimental results show that the average AUC value based on the multi-biological source information (MSCNE) model is 0.9630, which can reach higher performance than the other classic classifier, feature extractor mentioned and the other existing algorithms. The results show the MSCNE algorithm is effective to predict the correlation of miRNA-disease.

1 Introduction

Micrornas (miRNAs) are single-stranded non-coding small molecule RNAs with a size of about 23 nucleotides [1], which are processed by Dicer enzyme from a single-stranded RNA precursor with a hairpin structure of about 90 bases in size. Research and analysis of certain miRNAs have shown that miRNAs are involved in a series of important processes in life, including but not limited to the following conditions, early development, fat metabolism, cell proliferation, cell differentiation and cell death [2], [3], [4]. Besides, some studies shown that there is a significant correlation between the decline in the levels of two miRNAs and chronic lymphocytic leukemia, and that there may be a potential relationship between miRNAs and cancer [5]. Due to the extensiveness and diversity of miRNAs, it shows that miRNAs may have a very wide range of biological functions. Although the research on miRNA is still in infancy, with the continuous advancement of bioinformatics technology, a large number of miRNAs have been experimentally tested for genetic sequences. However, the cost of traditional bioinformatics technology is high, and it is difficult to identify the correlation between unknown miRNAs and diseases only through biological experiments. In recent years, the development of computers has been very rapid, and the application of computer technology can effectively predict the unknown correlation between miRNA and disease by using the correlation between biological information with low cost, which provides effective help for biologists to reduce their blindness [6], [7], [8].