HIV-1/HBV Coinfection Accurate Multitarget Prediction Using a Graph Neural Network-Based Ensemble Predicting Model

Area of Science:

• Computational biology
• Drug discovery
• Infectious diseases

Background:

• HIV and HBV coinfection pose significant global health challenges, affecting millions worldwide.
• Coinfection accelerates disease progression, increasing risks of liver disease and cancer.
• Current HIV treatment is complicated by drug interactions and hepatotoxicity, alongside HBV-related complications.

Purpose of the Study:

• To develop an efficient computational model for identifying multitargets for HIV-1 and HBV coinfections.
• To leverage machine learning for accelerating drug discovery in coinfected patients.
• To improve the accuracy of predicting potential drug targets for simultaneous HIV and HBV treatment.

Main Methods:

• Proposed a graph neural network (GNN)-based model for molecular feature extraction.
• Integrated an optimal supervised learner with the GNN to replace its output layer.
• Utilized a combination of Deep Molecular Particle Neural Network (DMPNN) and Gradient Boosting Decision Tree (GBDT) for prediction.

Main Results:

• The DMPNN + GBDT model demonstrated improved accuracy in binary-target predictions.
• The model efficiently identified potential multiple targets for HIV-1 and HBV simultaneously.
• This computational approach significantly accelerates the virtual screening of candidate drugs.

Conclusions:

• The developed GNN-based model offers a powerful tool for multitarget drug discovery in HIV/HBV coinfection.
• This approach can significantly reduce the time and cost associated with traditional drug development methods.
• The findings pave the way for more effective therapeutic strategies against HIV-1 and HBV coinfection.