Petri Nets with Fuzzy Logic (PNFL): reverse engineering and parametrization.

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• The DREAM4 challenge provided a realistic testbed for gene regulatory network inference.
• Inference methods faced challenging, noisy expression data, including time courses and various perturbations.

Purpose of the Study:

• To develop and evaluate a novel method for inferring gene regulatory networks.
• To assess the method's performance on complex, in silico biological networks.

Main Methods:

• Inferred and parameterized simulation models using Petri Nets with Fuzzy Logic (PNFL).
• Employed a completely automated approach for network reconstruction.

Main Results:

• PNFL achieved the best performance on DREAM4 in silico networks (size 10), with an 81% Area Under the Precision-Recall Curve (AUPR).
• Successfully reconstructed networks with cycles and oscillating motifs.
• Reliably distinguished activation/inhibition and dependent/independent regulation.
• Models performed well on unseen experimental conditions, like double knockout mutations.

Conclusions:

• PNFL offers a balance between expressive power and complexity for unified analysis of diverse biological datasets.
• The method provides intuitive graphical notation and colloquial fuzzy parameters.
• Effective inference of biological networks benefits from expressive and unified analytical methods.