Generating realistic in silico gene networks for performance assessment of reverse engineering methods

Area of Science:

• Systems Biology
• Bioinformatics
• Computational Biology

Background:

• Network inference algorithms require robust simulated data for accurate performance assessment.
• Existing random graph models inadequately represent the complex structures of real biological networks.

Purpose of the Study:

• To develop a method for generating biologically plausible in silico networks for realistic network inference algorithm evaluation.
• To provide a superior alternative to random graph models for simulating biological networks.

Main Methods:

• Network structures were generated by extracting modules from known biological interaction networks.
• The yeast transcriptional regulatory network was used as a test case to validate the method.
• The generated modules were assessed for biological plausibility, preserving functional and structural properties.

Main Results:

• Extracted modules from the yeast network demonstrated biologically plausible connectivity.
• The generated networks accurately reflect structural properties of real biological networks.
• The method was chosen for the "gold standard" networks in the DREAM 2008 challenge.

Conclusions:

• The proposed method effectively generates biologically plausible in silico networks.
• This approach enhances the reliability of assessing network inference algorithms.
• The technique offers a significant improvement over random graph models for network simulation.