Identifying disease candidate genes via large-scale gene network analysis.

Area of Science:

• Systems Biology
• Bioinformatics
• Genomics

Background:

• Gene Regulatory Networks (GRNs) are crucial for understanding complex biological systems and identifying disease-associated genes.
• Existing methods for GRN construction face challenges in handling uncertainty and integrating diverse biological data.

Purpose of the Study:

• To introduce Bayesian Model Averaging-based Networks (BMAnet), a novel reverse engineering technique for constructing large-scale GRNs.
• To evaluate BMAnet's performance in identifying disease candidate genes using network metrics like Random walk with restart (Rwr).

Main Methods:

• Developed BMAnet, an ensemble method integrating heterogeneous biological data to address model selection uncertainty.
• Employed network evaluation metrics, including Rwr, to assess identified networks.
• Applied BMAnet to brain tumor gene expression data (non-tumour, grade III, grade IV).

Main Results:

• BMAnet demonstrated superior performance compared to elastic-net and Gaussian graphical models in simulations.
• Analysis of brain tumor data identified 169 candidate genes.
• Identified candidate genes are associated with critical biological processes such as wound healing, apoptosis, and cell death.

Conclusions:

• BMAnet is a robust method for constructing GRNs and identifying disease candidate genes.
• The identified candidate genes offer insights into the molecular mechanisms underlying brain tumorigenesis.
• This approach facilitates the discovery of novel therapeutic targets for brain tumors.