Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

A Denoising Adversarial Model Based on Hyperellipsoidal Knowledge Representation Learning for DTI Prediction.

Chunming Yang, Fang Xiong, Yue Luo

IEEE Transactions on Computational Biology and Bioinformatics

|July 6, 2026

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Cucurbituril-based anion-conducting membranes with supramolecular nanopores.

Nature·2026

Same author

Mechanism analysis of uniconazole pretreatment improving waterlogging tolerance of different genotypes of Brassica napus L.

BMC plant biology·2026

Same author

Chiral hafnium phosphate catalyzed asymmetric de Mayo reaction via visible-light-induced dearomatization.

Nature communications·2026

Same author

Molecular Peripheral Electronegativity Modulates Crystallization Kinetics for Efficient Organic Solar Cells.

Small methods·2026

Same author

<i>Bacillus halotolerans</i> Strain 105 enhances wheat growth and suppresses diverse phytopathogens under saline conditions: an integrated genomic and phenotypic study.

Frontiers in microbiology·2026

Same author

Long noncoding RNA AFAP1-AS1 aggravates immunotherapy resistance in NSCLC by enhancing JAK2 and TYK2 translation.

Journal of advanced research·2026

This study introduces DAH-DTI, a novel model for predicting drug-target interactions (DTI) by addressing knowledge graph biases. DAH-DTI improves prediction accuracy by denoising data, generating quality samples, and using hyperellipsoid embedding.

Area of Science:

Bioinformatics
Computational Chemistry
Drug Discovery

Background:

Drug-target interaction (DTI) prediction is crucial for drug discovery.
Knowledge representation learning embeds drug knowledge graphs but often overlooks network biases like noise and sparsity.
These biases lead to incomplete semantic representations and suboptimal DTI prediction.

Purpose of the Study:

To propose a novel model, DAH-DTI, to address inherent biases in knowledge graph embedding for improved DTI prediction.
To enhance the accuracy and reliability of predicting potential drug-target interactions.

Main Methods:

Implemented a four-component model: knowledge graph denoising, high-quality negative sample generation, hyperellipsoid embedding using Mahalanobis distance, and DTI link prediction.

Related Experiment Videos

Denoised the knowledge graph to create a high-quality dataset and used adversarial negative sampling for pre-training.

Embedded tail entities into a hyperellipsoid for training the DTI prediction model.

Main Results:

DAH-DTI demonstrated significant improvements in accuracy (ACCU), recall (REC), mean reciprocal rank (MRR), and Hit@10 on a multi-level knowledge graph.
Effectively mitigated challenges posed by data noise, long-tail distribution, data sparsity, and complex relations.
Molecular docking simulations confirmed the reliability of DAH-DTI predictions for AR target interactions.

Conclusions:

DAH-DTI offers a robust approach to overcome inherent biases in network characteristics for knowledge graph embedding.
The model provides a promising strategy for accurate prediction of potential drug-target interactions, advancing drug discovery.
DAH-DTI enhances the semantic representation of knowledge graphs, leading to superior DTI prediction performance.