Combinatorial chromatin modification patterns in the human genome revealed by subspace clustering

Area of Science:

• Epigenetics and Genomics
• Computational Biology
• Molecular Biology

Background:

• Chromatin modifications regulate gene expression and support the 'histone code' hypothesis.
• Existing analytical methods limit the discovery of combinatorial histone modification patterns.
• Understanding these patterns is crucial for deciphering gene regulation.

Purpose of the Study:

• To develop a scalable algorithm for exhaustive identification of combinatorial chromatin modification patterns.
• To apply the algorithm to identify novel patterns in human epigenomes.
• To associate identified patterns with functional DNA elements.

Main Methods:

• Introduced CoSBI (coherent and shifted bicluster identification), a scalable subspace clustering algorithm.
• Applied CoSBI to analyze 39 genome-wide chromatin modification maps in human CD4(+) T cells.
• Performed performance comparisons to evaluate CoSBI's bicluster coherency and biological relevance.

Main Results:

• Identified 843 recurring combinatorial chromatin modification patterns (>0.1% of the genome).
• Observed 19 distinct chromatin modifications within these patterns, with 10 being highly prevalent.
• Discovered combinatorial modification signatures associated with eight classes of functional DNA elements.

Conclusions:

• CoSBI effectively identifies combinatorial modification patterns, overcoming limitations of previous methods.
• The discovered patterns provide insights into the 'histone code' and gene regulation mechanisms.
• Future applications of CoSBI to diverse epigenomes will enhance understanding of chromatin's role in gene expression regulation.