Inter-drug interaction recognition method based on deep learning

Abstract

The invention relates to an inter-drug interaction recognition algorithm based on deep learning, and relates to the field of medical big data, sequence data processing, graph data processing and deep learning. The method comprises the steps of 1) extracting substructure information and atomic information in a drug SMILES character string by a data preprocessing unit; 2) encoding the preprocessed drug information by using a multi-dimensional feature encoder, and encoding the drug information by using a sequence encoder and a graph encoder to obtain high-order representation of drug multi-dimensional features; and 3) inputting the multi-dimensional feature pairs of the drugs into an interaction recognition module, and recognizing the interaction between the drugs by using a feedforward neural network, including operations such as convolution, pooling and regularization. By using the trained model, a recognition result of interaction between drugs can be quickly given on an unknown interaction data set, and large-scale drug screening is facilitated. According to the method, the potential interaction between the drugs can be accurately recognized, and the effectiveness of drug combination can be favorably judged.

Description

technical field [0001] The method of the invention relates to the technical field of drug interaction recognition, in particular to a method for drug interaction recognition based on deep learning. Background technique [0002] Most human diseases are caused by complex biological processes, and it is difficult for us to completely cure them with any single drug. Patients often need to take more than one drug at the same time, an approach known as polydrug therapy. This type of treatment increases the frequency of drug-drug interactions and the possibility of adverse drug-drug reactions. Severe drug interactions may cause the drug to lose its original therapeutic effect, and may also lead to increased drug side effects, drug discontinuation, etc. Whether it is from the perspective of therapeutic effect or economic benefit, it is very important to identify unknown drug-drug interactions at an early stage. The traditional method of determining drug-drug interactions through ...

AI Technical Summary

Problems solved by technology

The traditional method of determining drug-drug interactions through clinical trials has problems such as high cost, long cycle, and low economic benefit

In this case, unknown drug-drug interactions pose a great challenge to drug safety

Images