Efficient prediction method for association relationship between circRNA and disease

Abstract

The invention discloses an efficient prediction method for a circRNA and disease association relationship. The method comprises the following steps: 1, downloading circRNA data and disease data from a database website; 2, calculating a circRNA Gaussian kernel similarity, a circRNA gene similarity, a circRNA sequence similarity, a disease Gaussian kernel similarity and a disease semantic similarity which are respectively a matrix CIS, a matrix CGS, a matrix CES, a matrix DIS and a matrix DSS; 3, constructing a circRNA comprehensive similarity matrix CS, and constructing a disease comprehensive similarity matrix DS; 4, obtaining similarity matrixes CRS and DRS by using a restart random walk algorithm; 5, respectively splicing the CRS, the DRS and the A, and performing feature extraction by using a PCA algorithm to obtain feature matrixes CF and DF; 5, constructing a heterogeneous adjacency matrix Acd according to CS, DS and the adjacency matrix A; constructing a heterogeneous feature matrix CD according to the CF and the DF; 6, finally, performing classification prediction on the Acd and the CD by using a graph convolutional neural network. The method provided by the invention is a brand-new method for predicting the association of circRNA and diseases.

Description

technical field [0001] The invention relates to the field of bioinformatics, in particular to a method for predicting the association between circRNA and disease. Background technique [0002] With the development of genomics and bioinformatics, especially the extensive application of high-throughput sequencing technology, scientists have discovered more and more non-protein-coding transcription units. In particular, circRNA, as the sponge adsorbent of microRNA, can indirectly regulate the expression of microRNA target genes and play an important role in the occurrence and development of human diseases. Therefore, circRNAs can be used as disease biomarkers and are widely used in disease diagnosis. [0003] At present, people have basically understood the formation and characteristics of circRNAs, but there are still many biological functions that are still unclear. Research on circRNA mainly focuses on its correlation with diseases. A large number of studies have shown tha...

AI Technical Summary

Problems solved by technology

[0005] However, the association between the vast majority of circRNAs and diseases is still unclear. On the other hand, studies have shown that the gene sequence of genetic RNAs also plays an important role in diseases, so it is useful to predict the association relationship between circRNAs, genes and sequences. Help study the relationship between circRNA and disease

However, predicting circRNA-disease associations based on experimental methods is expensive and time-consuming, and existing methods rarely use topological information of heterogeneous biological networks, or simply regard all objects as the same type, regardless of heterogeneity. Different subtle semantic meanings of different paths in the structural network, which will reduce the accuracy to a certain extent

Images