A bayesian framework that integrates heterogeneous data for inferring gene regulatory networks.

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Inferring gene regulatory networks (GRNs) from experimental data is crucial but challenging.
• mRNA expression profiles alone are insufficient for accurate GRN topology inference.
• Integrating diverse data sources can enhance GRN inference reliability.

Purpose of the Study:

• To develop a novel computational approach for more accurate GRN reconstruction.
• To integrate transcription factor binding site (TFBS) and protein-protein interaction (PPI) data into a Bayesian variable selection (BVS) algorithm.
• To infer GRNs from mRNA expression profiles under genetic perturbations.

Main Methods:

• A Bayesian variable selection (BVS) algorithm was developed.
• The BVS algorithm integrates TFBS and PPI data with mRNA expression profiles.
• The approach was applied to real experimental data and compared with LASSO regression methods.

Main Results:

• Integration of TFBS and PPI data significantly improved GRN accuracy compared to using expression profiles alone.
• The proposed BVS algorithm demonstrated superior performance in certain scenarios compared to LASSO regression-based methods.
• The study validates the utility of multi-data integration for robust GRN inference.

Conclusions:

• The proposed BVS approach effectively reconstructs GRNs by integrating TFBS and PPI data with expression profiles.
• This method offers a significant advancement in systems biology for understanding gene regulation.
• The findings highlight the potential of BVS for improving the accuracy and reliability of GRN inference.