HyperChIP

HyperChIP identifies genomic regions with hypervariable signals in ChIP-seq and ATAC-seq datasets across multiple samples to characterize epigenetic heterogeneity, including tumor heterogeneity and potential cancer subtypes.

Key Features:

• Statistical Framework: Employs scaled variances to account for the mean-variance dependence inherent in sequencing data and ranks genomic regions based on variability.
• Enhanced Statistical Power: Reduces the influence of true hypervariable regions during model fitting to increase the power for detecting and ranking variable regions.
• Tumor Heterogeneity Detection: Detects heterogeneity across tumor samples to facilitate exploration of epigenetic diversity and differences among cancer samples.

Scientific Applications:

• Cancer Genomics: Identifies hypervariable regions across tumor samples to investigate epigenetic contributions to cancer heterogeneity.
• Subtype Classification: Supports classification of potential cancer subtypes based on variability of ChIP-seq or ATAC-seq signals.
• Epigenetic Marker Discovery: Highlights regions with significant variability that may serve as candidate epigenetic markers associated with disease states.

Methodology:

Uses scaled variances to address the mean-variance relationship in sequencing data and fits a statistical model that minimizes bias from true hypervariable regions during model fitting to enable accurate ranking of variable genomic regions.