Integrative enrichment analysis: a new computational method to detect dysregulated pathways in heterogeneous samples

Area of Science:

• Bioinformatics
• Systems Biology
• Computational Biology

Background:

• Pathway enrichment analysis is crucial for understanding biological processes and disease mechanisms.
• Conventional methods often rely on differentially expressed genes (DEGs), which are less effective in heterogeneous disease samples.
• Genes with differential expression variance (DEVGs) offer valuable insights into specific biological system states, yet their role in pathway enrichment is challenging to measure.

Purpose of the Study:

• To introduce Integrative Enrichment Analysis (IEA), a novel method for pathway enrichment analysis that accounts for both DEGs and DEVGs.
• To develop a robust tool for analyzing biological pathways in complex, heterogeneous clinical contexts.
• To improve the identification of dysregulated pathways that are often underestimated by traditional methods.

Main Methods:

• IEA employs a novel enrichment measurement to quantify pathway status using both DEGs and DEVGs.
• IEA infers pathway crosstalks to associate identified dysregulated pathways with known target pathways.
• IEA recognizes subtype-factors by analyzing DEVGs' relative expressions, linking pathways to clinical indices.

Main Results:

• IEA effectively identifies dysregulated pathways containing both DEGs and DEVGs, outperforming other methods in human patient datasets.
• IEA successfully captures under-estimated dysregulated pathways in a proof-of-concept study on Diabetes.
• IEA-identified pathways show significant links to known disease pathways via crosstalk analysis, and identified subtype-factors correlate with genotype-phenotype associations.

Conclusions:

• IEA provides a new computational tool for enrichment analysis in complex clinical settings, addressing sample heterogeneity.
• IEA serves as a complementary and cooperative approach to conventional enrichment analysis methods.
• IEA enhances the ability to uncover biologically relevant pathways and potential disease subtypes.