cDREM: inferring dynamic combinatorial gene regulation.

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Genes are frequently regulated by combinations of transcription factors (TFs), crucial for development and stress response.
• Existing methods often overlook the dynamic and logical aspects of combinatorial TF regulation.
• There is a need for methods that integrate time-series data to capture regulatory dynamics.

Purpose of the Study:

• To introduce cDREM, a novel computational method for reconstructing dynamic models of combinatorial TF regulation.
• To identify TF combinations, their activation timing, and the logical functions governing gene expression.
• To improve the inference of combinatorial regulation by incorporating temporal dynamics.

Main Methods:

• cDREM integrates time-series gene expression data with static protein interaction data.
• The method employs a hidden Markov model framework.
• Sparse group Lasso is utilized to identify active TF subsets and their regulatory logic.

Main Results:

• cDREM successfully predicted combinatorial TF sets in yeast, aligning with genomic data and outperforming previous methods.
• Application to human flu response data identified known and novel combinatorial TF sets regulating immune responses.
• The method effectively captures the dynamics and logic of combinatorial gene regulation.

Conclusions:

• cDREM provides a powerful approach for modeling dynamic combinatorial gene regulation.
• The method enhances our ability to discover regulatory mechanisms underlying complex biological processes.
• cDREM facilitates the identification of key TF combinations involved in cellular responses.