Detection and identification of cis-regulatory elements using change-point and classification algorithms.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Transcriptional regulation involves factor binding to non-coding DNA regions.
• Identifying these binding regions is crucial for understanding tissue formation and gene therapy development.
• A universal biological code for characterizing binding regions is currently lacking, posing a challenge.

Purpose of the Study:

• To develop a computational method for identifying conserved binding sites in the human genome.
• To improve the understanding of transcriptional regulation through sequence analysis.
• To facilitate the interrogation of combined datasets for genomic research.

Main Methods:

• Extension of an alignment-based method (changept) to identify biologically significant clusters.
• Utilizing ontology and de novo motif analysis for cluster characterization.
• Application of a Bayesian method to combine binary classifiers for improved performance.

Main Results:

• Identification of clusters with biological significance using the extended changept method.
• Successful application of ontology and motif analysis to characterize identified clusters.
• Development of a composite classifier through Bayesian methods yielding enhanced performance.

Conclusions:

• The described computational method enables the identification of conserved binding sites in the human genome.
• This approach offers an alternative way to analyze combinations of existing datasets with alignment data.
• Facilitates a deeper understanding of transcriptional regulation and potential therapeutic targets.