Rank-statistics based enrichment-site prediction algorithm developed for chromatin immunoprecipitation on chip...

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• High-density oligonucleotide tiling arrays enable unbiased, genome-wide studies.
• Chromatin immunoprecipitation (ChIP) followed by array (ChIP-on-chip) maps in-vivo DNA-protein interactions.
• Identifying low-enrichment regions in non-canonical genomic areas presents a significant challenge.

Purpose of the Study:

• To develop a novel algorithm for identifying and statistically validating ChIP-enrichment regions.
• To optimize the detection of true positive low-enrichment sites.
• To provide a method applicable to general treatment-control experimental designs.

Main Methods:

• A rank and replicate statistics-based methodology was developed.
• The algorithm was optimized for identifying low but true positive enrichment signals.
• Sensitivity and specificity were validated using quantitative PCR and computational approaches.

Main Results:

• The algorithm demonstrates high concordance with independent biochemical validation methods.
• It effectively identifies true positive sites with low enrichment levels.
• Characterized sensitivity and specificity confirm algorithmic performance.

Conclusions:

• The algorithm ranks enrichment sites by intra- and inter-replicate consistency.
• Segmentation is achieved using meta p-values, array signal enrichment, or a composite measure.
• High sensitivities (88-95%) were achieved with different segmentation criteria.