ChromaSig

ChromaSig identifies chromatin signatures from tiling microarray and sequencing datasets using unsupervised learning to uncover functional genomic elements such as transcriptional promoters and enhancers.

Key Features:

• Unsupervised Learning Approach: Employs an unsupervised learning algorithm to analyze chromatin marks without prior labels or assumptions.
• Epigenetic Data Utilization: Processes epigenetic datasets, including histone modifications, from tiling microarray and sequencing experiments.
• Discovery of Chromatin Signatures: Detects clusters of distinct chromatin signatures that can correspond to regulatory elements such as transcriptional promoters and enhancers.
• Functional Classification: Classifies signatures into functional classes based on associations with transcription factors, coactivators, and evolutionary conservation.
• Genome-Wide Application: Has been applied to genome-scale datasets including a 1% sampling of the human genome in HeLa cells and a panel of 21 chromatin marks mapped genomewide by ChIP-Seq.

Scientific Applications:

• Genome Annotation: Enhances genome annotation by identifying genomic elements marked by characteristic chromatin modifications.
• Regulatory Element Discovery: Uncovers known and novel cis-regulatory elements not identifiable from genetic sequence alone.
• Functional Genomics Research: Supports classification of enhancers and other regulatory regions based on chromatin signature patterns.

Methodology:

Inputs tiling microarray and sequencing data capturing chromatin marks. Applies an unsupervised learning algorithm to detect patterns in epigenetic data. Groups identified signatures into clusters or classes based on shared characteristics. Compares identified signatures to known genomic elements and evolutionary conservation for validation and interpretation.