GBNet: deciphering regulatory rules in the co-regulated genes using a Gibbs sampler enhanced Bayesian network...

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Combinatorial regulation by transcription factors (TFs) governs complex gene expression patterns, especially in higher organisms.
• Understanding cooperative TF interactions is key to deciphering these intricate regulatory mechanisms.

Purpose of the Study:

• To introduce GBNet, a novel Bayesian network approach for identifying cooperative DNA motifs and sequence constraints involved in transcriptional regulation.
• To evaluate GBNet's performance against existing methods and demonstrate its utility in learning TF regulatory rules.

Main Methods:

• Bayesian network modeling
• Computational motif discovery
• Analysis of TF binding data

Main Results:

• GBNet demonstrated superior performance over other methods in simulated and yeast datasets.
• GBNet successfully identified regulatory rules between human TF YY1 and its co-factors, largely supported by existing literature.
• A specific spacing constraint between YY1 and E2F, identified by GBNet, was validated by independent TF binding experiments.

Conclusions:

• GBNet is an effective tool for deciphering the complex "grammar" of transcriptional regulation.
• The method aids in understanding combinatorial TF interactions and their impact on gene expression.