Drug relocation method based on machine learning

Abstract

The invention provides a drug relocation method based on machine learning, and relates to the technical field of machine learning. The method comprises the following steps: selecting a plurality of medicines as samples, and obtaining indications of each medicine; selecting multiple target protein data as drug sample features, and performing data dimension reduction on the drug-target protein vector by using a data dimension reduction algorithm based on machine learning; selecting a plurality of physicochemical characteristics of each medicine by using a correlation analysis algorithm; by taking the dimension-reduced drug-target protein vector features and drug physicochemical features as features of drug molecules and indications of drugs as labels, constructing a drug curative effect data set, establishing three gradient boosting trees, and training the three gradient boosting trees by using data in the drug curative effect data set. According to the invention, the three boosting trees are fused to establish a drug curative effect prediction model, a Kflood algorithm is utilized to perform multiple rounds of prediction on the curative effect of the N drugs, and finally m drugs effective for treating a certain disease are predicted.

Description

technical field [0001] The present invention relates to the technical field of machine learning, in particular to a drug repositioning method based on machine learning. Background technique [0002] According to statistics, it takes about 15 years for a new drug to be conceived, synthesized as a lead compound, passed through clinical trials, and finally successfully marketed, costing about US$1 billion. And this cost is increasing year by year. Moreover, there are risks in drug research and development. If problems with the drug are found in the later stage of research and development, the previous investment will be wasted, so the risk is very high. [0003] Drug repositioning refers to the discovery of new indications for marketed drugs, and is an important application field of network pharmacology. The drug repositioning strategy is one of the strategies with the best risk-to-benefit ratio in the currently known drug development strategies, and it is also one of the eff...

AI Technical Summary

Problems solved by technology

[0005] Drug repositioning has the characteristics of high efficiency and low cost. Since the outbreak of the new crown epidemic, how to screen out new crown drugs has become an urgent problem to be solved. However, the traditional drug development cycle is too long, which requires a lot of time and manpower and material resources financial resources

Images