Identifying disease network perturbations through regression on gene expression and pathway topology analysis

Area of Science:

• Systems Biology
• Bioinformatics
• Computational Biology

Background:

• Network-based approaches are crucial for studying complex diseases in Systems Biology.
• Detecting network perturbations that disrupt biological functions due to disease is a significant challenge.

Purpose of the Study:

• To introduce a novel network-based pathway analysis method.
• To isolate causal interactions with significant regulatory roles in disease-perturbed pathways.

Main Methods:

• Utilized gene expression data and Random Forest regression models.
• Assessed gene interactivity strengths within disease-perturbed networks.
• Incorporated KEGG pathway maps for prior knowledge of pathway topology.

Main Results:

• Developed a method to output a network reflecting disease-induced perturbations.
• Demonstrated efficacy on a serous papillary ovarian cancer experiment.
• Highlighted functional roles of high-impact interactions and key gene regulators.

Conclusions:

• The method effectively identifies critical perturbations and regulators in disease networks.
• Results align with experimentally validated knowledge, advancing understanding of disease mechanisms.