IncRNA and disease association prediction method fusing heterogeneous network and graph neural network

Abstract

The invention relates to the field of data mining in bioinformatics, in particular to an lncRNA and disease association prediction method fusing a heterogeneous network and a graph neural network. The method mainly comprises the following steps: (1) collecting related data; (2) calculating the semantic similarity of the disease, the target similarity of the disease, the sequence similarity of the lncRNA and the functional similarity of the lncRNA; (3) constructing a heterogeneous network net1 by using DDSsem, LLSfun, LDA, LMA and DMA; and constructing a heterogeneous network net2 by using the DDStar, the LLSseq, the LDA, the LMA and the DMS; (4) constructing a neural network model with an attention mechanism, extracting topological structure features in the network by an encoder part through GCN, and fusing the features between nodes, between graphs and between layers by using the attention mechanism; (5) constructing and training a BP neural network; (6) predicting by using the trained BP neural network; and (7) performing an experiment to verify the performance of the prediction model.

Description

technical field [0001] The invention relates to the field of data mining in bioinformatics, in particular to a lncRNA-disease association prediction method based on heterogeneous graphs and graph deep learning of fusion multidimensional data. Background technique [0002] With the completion of the Human Genome Project, life science research has entered the era of functional genomics. Its task is to annotate genome functions, grasp the role of gene products in life activities, and establish the relationship between genes and diseases. Noncoding RNAs have long been viewed as recorded noise, devoid of any biological noise. However, more and more studies have shown that non-coding RNA plays an important role in many biological processes, and its function involves almost all biological processes of organism physiology and pathology. lncRNA plays an important role in the metastasis and development of various diseases. Therefore, in-depth research and exploration of the relations...

AI Technical Summary

Problems solved by technology

[0005] At present, computational methods to identify disease-related lncRNAs have attracted the attention of many scholars. Researchers have developed many related computational models and accumulated a large amount of data, but they still face many challenges

Images