KEGG2Net: Deducing gene interaction networks and acyclic graphs from KEGG pathways.

Area of Science:

• Bioinformatics
• Systems Biology
• Computational Biology

Background:

• The Kyoto Encyclopedia of Genes and Genomes (KEGG) database offers comprehensive biological pathway information.
• Omics data analysis often requires gene-centric network representations, focusing on gene-gene interactions.
• Bayesian networks and other methodologies necessitate directed acyclic graph (DAG) formats.

Purpose of the Study:

• To develop KEGG2Net, a web resource for generating gene-gene interaction networks from KEGG pathways.
• To enable the conversion of these networks into DAGs by removing cycles.
• To evaluate different cycle removal methods for their efficacy in representing biological networks.

Main Methods:

• KEGG pathway data was utilized to construct gene-gene interaction networks.
• Four distinct methods were implemented for cycle removal to generate DAGs.
• Synthetic gene expression data was generated to assess the performance of the DAGs.
• Comparative analysis focused on DAG fitness to data and edge reduction.

Main Results:

• KEGG2Net successfully generates gene-gene interaction networks and corresponding DAGs from KEGG pathways.
• An ensemble method for cycle removal demonstrated superior performance in converting networks to DAGs.
• The generated networks and DAGs are available in multiple formats for diverse applications.

Conclusions:

• KEGG2Net provides a valuable tool for converting KEGG pathways into gene-centric networks and DAGs.
• The ensemble cycle removal method is recommended for creating robust DAGs from pathway data.
• The resource facilitates advanced analysis of omics data by offering interpretable gene interaction networks.