CCAT

CCAT applies a linear signal-noise model and control-based statistical strategies to ChIP-seq data to detect and assess the significance of strong and weak ChIP-enriched sites.

Key Features:

• Linear Signal-Noise Model: Employs a linear signal-noise model with a noise rate parameter to quantify background in ChIP libraries and improve detection of weak ChIP signals.
• Iterative Noise Rate Estimation: Uses an iterative algorithm to estimate the noise rate from control library data.
• Library-Swapping Strategy for FDR Estimation: Implements a library-swapping strategy to estimate false discovery rate (FDR) for ChIP-enriched site calls.
• General-Purpose Framework: Operates as a general-purpose framework applicable to diverse ChIP-seq datasets.

Scientific Applications:

• Histone modification analysis (H3K4me3 and H3K36me3): Demonstrated increased sensitivity in predicting ChIP-enriched sites, identifying a greater number of enriched sites including weak signals to inform chromatin feature interpretation.

Methodology:

Linear signal-noise model with a noise rate parameter; iterative noise rate estimation using control library data; and library-swapping FDR estimation integrated within a unified framework.