Interpretable adverse drug reaction discovery method based on literature knowledge graph

Abstract

The invention discloses an interpretable adverse drug reaction discovery method based on a literature knowledge graph. The method comprises the following steps: S1, extracting four entities from medical literature data: a disease, a biomarker, a drug and an adverse reaction; s2, constructing the literature knowledge graph by using the four entities and utilizing a naive Bayesian model: the literature knowledge graph comprises vertexes and edges, the vertexes comprise vertexes of four entity types, the edges represent the relationship between the two vertexes, the weight on the edges represents the correlation between the two vertexes, and the correlation is described through importance indexes; s3, based on the adverse drug reaction in the literature knowledge map, comparing a drug specification, and discovering potential adverse reactions; and S4, providing reasonable biomarker path interpretation for the potential adverse reaction based on a literature knowledge graph. Potential adverse reactions of drugs are explored through a technical means of computational medicine, and power is provided for mechanism research of the adverse reactions.

Description

technical field [0001] The invention relates to the technical fields of drug informatics and bioinformatics, in particular to an interpretable method for discovering adverse drug reactions based on a document knowledge map. Background technique [0002] Adverse drug reactions are a cause of severe morbidity and mortality for patients and a source of economic burden to healthcare systems. At present, drug-induced diseases have become the fifth leading cause of death, and about 1 / 3 of the global deaths are caused by improper use of therapeutic drugs. Especially for cancer patients, they have a relatively high incidence of adverse drug reactions when using anticancer drugs, and are more likely to experience rare and serious adverse reactions, which will seriously affect their quality of life. However, due to the limited sample size and generalizability of clinical trials, the identification of rare and serious adverse reactions before marketing is limited, so exploring potenti...

AI Technical Summary

Problems solved by technology

[0004] Based on the above problems, the present invention provides an interpretable method for discovering adverse drug reactions based on literature knowledge graphs, which is dedicated to exploring potential adverse drug reactions through computational medicine techniques and solving the problem that potential adverse drug reactions are not easy to identify

Images