PaCMAP-embedded convolutional neural network for multi-omics data integration.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Multi-omics data integration is crucial for advancing predictive models in healthcare.
• Precision in cancer diagnosis is vital for effective clinical decision-making.
• Current prognostic methodologies require enhancement for improved accuracy.

Purpose of the Study:

• To develop an innovative methodology for integrating copy number alteration (CNA), DNA methylation, and gene expression data.
• To improve prognostic accuracy in cancer diagnosis.
• To enhance clinical decision-making for prostate cancer patients.

Main Methods:

• Utilized PaCMAP for dimensionality reduction to merge three omics data into a 2D map.
• Employed an RGB coloring scheme to visualize the integrated multi-omics data.
• Applied a convolutional neural network (CNN) to the 2D maps for Gleason score prediction.

Main Results:

• The proposed model achieved a prediction accuracy of 98.89%.
• The model demonstrated a high Area Under the Curve (AUC) of 0.9996.
• Outperformed the existing i-SOM-GSN model in predicting Gleason scores.

Conclusions:

• Multi-omics data integration significantly enhances the prediction of health outcomes.
• The combined approach of PaCMAP, RGB visualization, and CNN provides a robust framework for heterogeneous omics data integration.
• Findings support the advancement of personalized medicine and clinical decision support for prostate cancer.