Identification of Pandemic Risk for Avian Influenza Virus With Graph Cross Attention Networks.

Area of Science:

• Virology
• Computational Biology
• Epidemiology

Background:

• Influenza A virus presents a significant global health risk due to its rapid mutation and pandemic potential.
• Accurate pandemic risk prediction is crucial for public health and societal stability.
• Current deep learning methods primarily rely on viral genome sequences, neglecting crucial genetic reassortment dynamics.

Purpose of the Study:

• To develop an advanced method for predicting influenza virus pandemic risk.
• To incorporate viral gene flow, specifically genetic reassortment, into pandemic risk prediction models.
• To improve the accuracy and effectiveness of influenza pandemic forecasting.

Main Methods:

• Proposed Graph Cross Attention Networks for influenza pandemic identification (GCAN-Flu).
• Constructed an influenza virus gene flow network based on genetic reassortment relationships.
• Utilized Graph Cross Attention (GCA) layers to integrate viral genome sequences with gene flow information for multimodal representation.
• Predicted pandemic risks using these multimodal representations.

Main Results:

• GCAN-Flu demonstrated superior performance compared to existing sequence-based deep learning methods.
• The integration of gene flow information significantly enhanced pandemic risk prediction accuracy.
• The model provides a powerful new tool for influenza pandemic forecasting.

Conclusions:

• GCAN-Flu offers a more comprehensive approach to influenza pandemic risk assessment by including genetic reassortment.
• This method advances the field of computational virology and pandemic preparedness.
• The developed model shows promise for early detection and mitigation of future influenza pandemics.