A Drug-Target Interaction Prediction Method Based on Contrastive Learning of Supervised Synergy Graphs

A prediction method and supervised technology, applied in the field of drug-target relationship prediction, can solve the problems of too many complicated steps in the model and tedious feature extraction, and achieve the effect of enhancing learning ability, reducing processing steps, and high prediction accuracy.

Active Publication Date: 2022-08-09
NORTHEAST FORESTRY UNIVERSITY
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem that the traditional machine learning method needs to rely on cumbersome manual feature extraction and the model has too many complicated steps, and proposes a drug-target interaction prediction method based on supervised synergy graph comparison learning

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Drug-Target Interaction Prediction Method Based on Contrastive Learning of Supervised Synergy Graphs
  • A Drug-Target Interaction Prediction Method Based on Contrastive Learning of Supervised Synergy Graphs
  • A Drug-Target Interaction Prediction Method Based on Contrastive Learning of Supervised Synergy Graphs

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0046] Specific implementation mode 1. Combination figure 1 This embodiment will be described. A method for predicting drug-target interaction based on supervised synergy graph comparative learning described in this embodiment specifically includes the following steps:

[0047] Step S1, extracting drug information, protein information, disease information and drug side effect information from the database, and constructing a drug heterogeneous information network and a protein heterogeneous information network according to the extracted information;

[0048] Step S2, obtaining the final representation of each drug in the drug heterogeneous information network based on the first encoder, and obtaining the final representation of each protein in the protein heterogeneous information network based on the second encoder;

[0049] The specific process of obtaining the final representation of each drug in the drug heterogeneous information network based on the first encoder is as f...

specific Embodiment approach 2

[0063] Embodiment 2: The difference between this embodiment and Embodiment 1 is that the specific process of step S1 is:

[0064] Extracting drug information from the DrugBank database, the drug information includes drug-drug interaction information and known drug-protein interaction information;

[0065] Extract protein information from the HPRD database, where the protein information is protein-protein interaction information;

[0066] extracting disease information from a toxicogenomics database, the disease information including disease-drug relationship information and disease-protein relationship information;

[0067] Extract drug side effect information from the SIDER database, where the drug side effect information is the relationship information between drugs and side effects;

[0068] Construct a drug heterogeneous information network according to the interaction information between drugs, the interaction information between drugs and proteins, the relationship info...

specific Embodiment approach 3

[0074] Embodiment 3: The difference between this embodiment and Embodiment 1 or 2 is that the attention mechanism is used to give different weights to different meta-paths, and the representations of each meta-path are weighted and summed according to the weights, and the weighted and the result as the final representation of the drug corresponding to this node; it is specifically:

[0075]

[0076] in, is the representation of the i-th metapath output by the first encoder, W is the weight matrix, b is the bias vector, q is the transformation vector, is the weight given to the i-th meta-path;

[0077] Weighted summation of the representations of individual metapaths:

[0078]

[0079] where h drug is the final representation of the drug corresponding to this node, i=1,2,...M d , M d is the total number of all meta-paths including this node in the drug heterogeneous information network.

[0080] Other steps and parameters are the same as in the first or second emb...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

A drug-target interaction prediction method based on supervised synergy graph contrastive learning belongs to the technical field of drug-target relationship prediction. The invention solves the problems that the traditional machine learning method needs to rely on tedious manual feature extraction and the model has too many complicated steps. The drug-target interaction prediction method of the present invention uses graph comparison learning to enhance the learning ability of the model. In the whole prediction process, no manual operation is required, that is, it does not rely on cumbersome manual feature extraction, and the end-to-end idea is applied to reduce The processing steps of the model are reduced, the complexity of the model is reduced, and a high prediction accuracy is ensured. It is obtained through experiments that the area under the Roc curve of the prediction method of the present invention can reach 0.9764, and the area under the PR curve can reach 0.9761. The present invention can be used to predict drug-target relationships.

Description

technical field [0001] The invention belongs to the technical field of drug-target relationship prediction, in particular to a drug-target interaction prediction method based on supervised synergy graph comparative learning. Background technique [0002] We refer to molecules that can bind to drugs and play a specific role in cells as drug targets, and proteins are the main drug targets. In the process of drug discovery and redirection, we put our minds to countless tests and experiments to find safe and effective compounds as medicines. So, drug discovery is a difficult process. Furthermore, identifying interactions between drugs and protein targets is not only a critical step in drug discovery, but also provides guidance for drug repositioning, multidrug pharmacology, drug resistance, and side effect prediction. [0003] In recent years, with the development of computer technology, many researchers use computer technology to calculate the probability of drug-target inter...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G16H70/40G16B40/00G16B5/00G06N3/04G06N3/08
CPCG16H70/40G16B5/00G16B40/00G06N3/08G06N3/045Y02A90/10
Inventor 汪国华李洋乔冠宇
Owner NORTHEAST FORESTRY UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap