Research on a novel gene sequence prediction and homomorphic encryption method based on Mamba-VMD

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Gene sequence prediction is vital for understanding gene function and evolution.
• Transmitting sensitive gene sequences in plaintext to cloud environments risks privacy.
• Existing methods lack robust security for cloud-based bioinformatics data analysis.

Purpose of the Study:

• To develop a secure and accurate method for gene sequence prediction in cloud environments.
• To integrate Mamba neural networks with homomorphic encryption for privacy-preserving analysis.
• To validate the proposed method using monkeypox virus gene sequence data.

Main Methods:

• Gene sequence prediction utilizing the Mamba neural network.
• VMD modal decomposition for preprocessing gene sequences.
• Homomorphic encryption (CKKS) for secure cloud data processing.
• Spatial similarity analysis on encrypted data.

Main Results:

• Achieved low prediction errors (MAE: 0.0140, MSE: 0.0003, RMSE: 0.0190, MAPE: 0.6527, MSPE: 789.52959).
• Demonstrated a low average CKKS homomorphic encryption computation error (0.582492456).
• Successfully enabled secure similarity calculations for gene sequence data in the cloud.

Conclusions:

• The combined Mamba neural network and homomorphic encryption approach ensures data privacy.
• The method provides accurate gene sequence prediction while maintaining security.
• This framework is suitable for sensitive bioinformatics data analysis in cloud settings.