Medicament combination predicting method based on medicament similarity network data

Abstract

The invention discloses a medicament combination predicting algorithm based on a medicament similarity network, wherein the algorithm mainly settles a problem of low accuracy of a medicament combination predicting result in prior art. The medicament combination predicting algorithm is characterized by comprising steps of downloading data from a medicament database, constructing the medicament similarity network; downloading data from a medicament combination database; obtaining a network importance score based on medicament combination data and the medicament similarity network; obtaining a total set of random walk paths of multiple networks by means of a biased random walk algorithm, and training the total set of the random walk paths for obtaining a medicament node characteristic vector; training the medicament combination data and the medicament node characteristic vector by means of a random forest model, and obtaining a medicament combination predicting model; and inputting medicament pair data into the medicament combination predicting model, obtaining a medicament pair score, and using the medicament pair with score which is higher than 0.5 as a cooperation medicament combination. The medicament combination predicting algorithm improves medicament combination predicting accuracy and can be used for predicting the cooperative medicament combination candidate.

Description

technical field [0001] The invention belongs to the technical field of bioinformatics, and in particular relates to a drug combination prediction method, which can be used to provide candidate drug combination references for drug combination experiments. Background technique [0002] Drug combination therapy refers to the treatment of disease patients with two or more drugs at the same time. According to the effect of drug combination therapy, the relationship of drug combination therapy can be divided into three categories: the first category, the simultaneous use of two drugs at a low dose can achieve the effect of a single drug at a high dose, such a drug combination relationship is called It is synergy. In the second category, taking two drugs at the same time leads to strong side effects or less therapeutic effect than taking a single drug. Such a drug combination relationship is called antagonism. In the third category, the effects of two drugs taken together and tak...

AI Technical Summary

Problems solved by technology

In the second category, taking two drugs at the same time leads to strong side effects or less therapeutic effect than taking a single drug. Such a drug combination relationship is called antagonism

This type of method makes good use of the multi-omics data of drugs, but the drug combination features constructed by it do not reflect the topological information of the drug similarity network, and the constructed drug combination feature dimensions are fixed, and the accuracy of the obtained results is low.

Images