WordCluster

WordCluster identifies clusters of DNA words (k-mers) and other genomic elements by assessing statistical significance of distances between consecutive copies to detect non-random spatial aggregations.

Key Features:

• Statistical significance: Analyzes distances between consecutive k-mers or genomic elements and assigns statistical significance to detected clusters.
• Distance-based framework: Implements a distance-based approach that moves beyond density-based and sliding-window methods that rely on arbitrary distance thresholds.
• Co-localization with gene annotations: Determines co-localization of identified clusters with gene annotations to connect clusters to nearby genes.
• Annotation enrichment: Performs annotation enrichment analysis for genes overlapping identified clusters to support functional interpretation.

Scientific Applications:

• Detection of cytosine methylation patterns: Identifies clusters of CAG/CTG dinucleotides and reveals differences in cytosine methylation between clustered and non-clustered regions.
• Olfactory receptor gene clustering: Detects statistically significant clusters of olfactory receptor (OR) genes in the human genome to reveal biologically meaningful groupings.

Methodology:

Calculates distances between consecutive copies of k-mers or other genomic entities and assigns statistical significance to distinguish true clusters from random distributions, contrasting with density-based and sliding-window techniques.