Drug-disease association prediction method and system

Abstract

Aiming at the limitation of the existing drug relocation technology, the invention provides a drug-disease association prediction method and system, and the method partially reduces the sparsity of an existing drug-disease original association matrix by applying a weight K nearest neighbor algorithm; meanwhile, a linear feature extraction mode and a nonlinear feature extraction mode are integrated when the similarity feature information of the medicine and the disease is extracted, so that more comprehensive similar feature information is obtained, and the relationship between the medicine and the disease can be fully reflected; and a bipartite graph diffusion method is adopted to calculate a drug-disease associated prediction score, so that excellent prediction performance is obtained.

Description

technical field [0001] The invention relates to the technical field of drug repositioning, in particular to a drug-disease association prediction method and system based on integrated similarity and bipartite graph diffusion algorithm. Background technique [0002] Drug repurposing aims to identify new indications for existing drugs, which will greatly reduce the cost and time of drug development. In recent years, researchers have explored a variety of computational methods for mining the link between existing drugs and diseases, including methods based on machine learning methods, matrix factorization, networks and graphs. However, all the methods applied to drug repositioning so far directly use the information of the unprocessed original drug-disease association matrix. Since only a small part of these similar association information has been proved and recorded in actual situations, there is still There is a lot of similar correlation information that needs to be proven...

AI Technical Summary

Problems solved by technology

However, all the methods applied to drug repositioning so far directly use the information of the unprocessed original drug-disease association matrix. Since only a small part of these similar association information has been proved and recorded in actual situations, there is still There is a lot of similar correlation information that needs to be proved urgently, resulting in the sparseness of the original correlation matrix data used in the prediction method

In addition, for the extraction of similarity features between drugs and diseases, there is still a lack of sufficient mining methods

[0003] The publication date is 2021.03.26, and the announcement number is CN112562795A Chinese patent application: based on the fusion of multiple similarities The drug similarity is calculated from the target protein and drug side effects data, and then weighted and summed to obtain the fused drug similarity, and the drug-disease correlation prediction value is calculated based on the fused drug similarity, so as to avoid relying on a single data source to calculate the drug-to-disease prediction value deviation problem; it uses the Tanimoto coefficient to calculate the disease similarity, and the drug-disease association prediction value calculated based on the disease similarity is weighted and summed with the drug-disease association prediction value calculated based on the drug similarity to obtain the fused The predictive value of drugs on diseases, in order to improve the reliability of the predictive value; broadening the data source can certainly avoid the problem of prediction bias to a certain extent, but the program has too high requirements on the data source, and it is more complicated to implement, so there are still certain limitations

Images