Effective knowledge graph embeddings based on multidirectional semantics relations for polypharmacy side effects prediction | JoVE Visualize

Area of Science:

Pharmacology
Bioinformatics
Computational Biology

Background:

Polypharmacy, the concurrent use of multiple drugs, presents significant challenges in healthcare due to increased risk of adverse drug events and mortality.
Traditional machine learning methods for predicting drug side effects require extensive feature engineering, while existing knowledge graph approaches inadequately model complex drug-related interactions.
Accurate prediction of polypharmacy side effects is crucial for patient safety and effective disease management.

Purpose of the Study:

To develop a novel model that effectively incorporates complex side effect relations into knowledge graph embeddings for improved polypharmacy safety.
To address the limitations of existing methods in modeling intricate drug interactions and reduce the computational burden of feature extraction.
To enhance the prediction accuracy of adverse drug events associated with polypharmacy.

Main Methods:

Proposed a novel model integrating complex side effect relationships within knowledge graph embeddings.
The model facilitates multidirectional semantic translation and transmission with reduced parameter count, enhancing scalability for large knowledge graphs.
Utilized knowledge graph embeddings to capture intricate drug-drug interactions and their associated side effects.

Main Results:

The proposed model demonstrated superior performance compared to state-of-the-art methods in predicting polypharmacy side effects.
Achieved higher accuracy as measured by the average area under the ROC and precision-recall curves.
The model's ability to handle complex relations led to improved prediction outcomes.

Conclusions:

The novel knowledge graph embedding model effectively captures complex side effect relations, outperforming existing methods.
The approach offers improved scalability and efficiency for predicting polypharmacy-related adverse events.
This advancement holds significant potential for enhancing patient safety in polypharmacy management.