A scale-free structure prior for graphical models with applications in functional genomics.

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Gene regulatory network reconstruction is crucial for understanding cellular mechanisms.
• Graphical models are widely used for inferring these networks from gene expression data.
• Current methods often use random structure priors, which may not reflect the scale-free nature of biological networks.

Purpose of the Study:

• To introduce and evaluate a novel scale-free structure prior for graphical models in gene network inference.
• To improve the accuracy of reconstructing biological networks, particularly those with scale-free properties.
• To identify potential hub genes in complex diseases like breast cancer.

Main Methods:

• Developed a scale-free structure prior based on scale-free network models.
• Designed a Metropolis-Hastings sampler incorporating node labeling for network inference.
• Applied the prior and sampler to simulated data and real-world gene expression data from breast cancer studies.

Main Results:

• The scale-free structure prior demonstrated superior performance in recovering scale-free networks compared to random structure priors.
• The method successfully identified known and novel hub genes, including SLC39A6, a zinc transporter implicated in breast cancer metastasis.
• The prior aids in identifying candidate hub genes, guiding future biological research and experiments.

Conclusions:

• The scale-free structure prior is a valuable tool for enhancing gene network inference, especially for scale-free biological networks.
• This approach can effectively identify critical regulatory genes (hubs) relevant to disease mechanisms.
• The findings support the use of this prior to direct molecular biology hypotheses and experimental design.