KFDAE: CircRNA-Disease Associations Prediction Based on Kernel Fusion and Deep Auto-Encoder

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Circular RNAs (circRNAs) are crucial in disease diagnosis and treatment.
• Wet-lab experiments for verifying circRNA-disease associations are costly and time-consuming.
• Existing computational methods often fail to fully leverage attribute dependencies.

Purpose of the Study:

• To develop a novel computational method for predicting circRNA-disease associations (CDAs).
• To address limitations of existing methods in handling attribute dependencies and limited verified associations.
• To provide a more efficient and cost-effective approach compared to wet-lab validation.

Main Methods:

• Kernel Fusion and Deep Auto-encoder (KFDAE) method.
• Non-linear fusion of circRNA and disease similarity kernels.
• Deep auto-encoder for feature extraction and dimensionality reduction.
• Three-layer deep feedforward neural network for prediction score generation.

Main Results:

• KFDAE demonstrated superior performance compared to existing computational methods.
• Experimental results validated the effectiveness and practical significance of KFDAE.
• The method successfully captured comprehensive information for credible CDA prediction.

Conclusions:

• KFDAE offers a powerful computational tool for predicting circRNA-disease associations.
• The method provides a valuable resource for identifying potential CDAs for further wet-lab investigation.
• KFDAE enhances the efficiency and accuracy of disease diagnosis and treatment strategies involving circRNAs.