Drug oral availability and toxicity prediction method based on graph convolutional neural network

Abstract

The invention discloses a graph convolutional neural network-based drug oral availability and toxicity prediction method. The method comprises the steps of S1, preparing an initial training set; s2, establishing a graph model of drugs, and obtaining a training set; s3, training a graph convolutional neural network and a full-connection neural network by using the training set, and fitting a molecular descriptor of the drug and a mapping relationship between a graph model and oral availability and toxicity of the drug; s4, performing numerical modification on each molecular descriptor feature in the training data, predicting the modified training data by using a neural network, and determining a corresponding predicted value error; s5, sorting all the molecular descriptor features of the medicine, calibrating the molecular descriptor features located in the preorder, deleting the molecular descriptor features of the medicine which are not calibrated, and updating the training data; and S6, retraining the graph convolutional neural network and the full-connection neural network constructed in the step S3. According to the method, the drug oral availability and toxicity prediction model with high prediction precision can be obtained.

Description

technical field [0001] The invention relates to the technical field of computer-aided drug design, in particular to a method for predicting the oral availability and toxicity of a drug based on a graph convolutional neural network. Background technique [0002] In the drug screening stage, after obtaining a drug with specific pharmacological properties, its efficacy and safety need to be evaluated. The traditional method uses clinical pharmacology to carry out animal tests and Phase I, II, and III clinical trials of new drugs to evaluate their safety and effectiveness, and conduct Phase IV clinical trials in the application stage after the drug is launched to investigate the efficacy and effectiveness of drugs. Adverse reactions. Due to the limited number of patients participating in clinical trials, and a large number of new drugs are put into clinical trials every year, it takes a lot of manpower, material resources and time costs for a new drug to be developed from devel...

AI Technical Summary

Problems solved by technology

However, in the existing prediction methods using deep learning networks, only the molecular descriptors or molecular fingerprint features of the drug are considered, and the structural information of the molecule is lacking when the features are input into the subsequent prediction model, and the molecular descriptor features are not screened. , when all molecular descriptor features are used, it will lead to a decrease in the prediction accuracy of drug availability and toxicity

Images