An interpolated Markov model polishes Gibbs sampling's ability in detecting regulatory elements

Area of Science:

• Genomics and Bioinformatics
• Molecular Biology
• Computational Biology

Background:

• Microarray analysis identifies coregulated genes via coexpression profiles.
• Investigating cis-acting regulatory elements in upstream sequences is crucial for understanding gene transcription.
• Coexpressed genes are hypothesized to share regulatory elements.

Purpose of the Study:

• To develop and evaluate a modified Gibbs sampling algorithm for detecting regulatory elements.
• To identify cis-acting regulatory elements in the upstream regions of coexpressed genes.
• To analyze upstream sequences of coexpressed genes in Saccharomyces cerevisiae.

Main Methods:

• A modified Gibbs sampling algorithm incorporating an interpolated Markov model (IMM) was developed.
• The algorithm was tested using simulated data to assess its performance.
• The algorithm was applied to analyze upstream sequences of coexpressed gene clusters in yeast.

Main Results:

• The improved Gibbs sampling algorithm demonstrated superior performance in extracting less-conserved elements compared to existing methods.
• The algorithm outperformed single nucleotide independent models and fixed higher-order Markov models.
• Several putative regulatory motifs were identified in the upstream sequences of coexpressed gene clusters in Saccharomyces cerevisiae.

Conclusions:

• The enhanced Gibbs sampling algorithm is effective for identifying gene regulatory elements, particularly less-conserved ones.
• The findings contribute to a better understanding of transcriptional regulation in coexpressed gene networks.
• The identified motifs offer insights into regulatory pathways in Saccharomyces cerevisiae under diauxic shift conditions.