tensorGSEA: Detecting Differential Pathways in Type 2 Diabetes via Tensor-Based Data Reconstruction.

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Detecting significant signaling pathways is crucial for understanding complex disease mechanisms.
• Tensor decomposition offers a powerful approach for multi-dimensional data analysis and reconstruction.

Purpose of the Study:

• To introduce tensorGSEA, a novel tensor-based gene set enrichment analysis method.
• To identify differential pathways involved in disease progression using a data reconstruction approach.

Main Methods:

• Gene expression profiles were organized into three-dimensional tensors (genes, samples, periods).
• Tensors were compressed into lower-rank core tensors for data reconstruction.
• Cross-state data reconstruction between control and disease groups identified differential pathways.

Main Results:

• Identified critical pathways with diabetes-specific functions not found by other methods.
• Demonstrated the efficiency of tensorGSEA in evaluating pathway statistical significance.
• Selected pathways showed potential as biomarkers for diabetic state classification.

Conclusions:

• TensorGSEA effectively identifies significant pathways in disease development.
• The identified pathways can serve as diagnostic biomarkers for diseases like diabetes.
• This method offers a novel approach to analyzing complex biological data.