mosaics

mosaics models ChIP-seq background and detects enrichment peaks using the MOSAiCS-HMM statistical framework for one-sample and two-sample analyses to enable genome-wide profiling of transcription factor binding sites and histone modifications.

Key Features:

• MOSAiCS-HMM statistical framework: Implements a Hidden Markov Model–based statistical framework (MOSAiCS-HMM) for ChIP-seq analysis in both one-sample and two-sample settings.
• Background model: Incorporates a background model that accounts for mappability and GC content to explain background signal distribution in ChIP-seq data.
• Flexible mixture model: Uses a flexible mixture model to detect enrichment peaks across one-sample and two-sample analyses.
• Bias modeling: Explicitly models biases arising from ChIP-seq technology and standard pre-processing protocols.
• Naked DNA sequencing evaluation: Employs naked DNA sequencing experiments (deproteinized, sheared DNA) to characterize and validate background distributions.
• Data fitting: Fits observed ChIP-seq data to improve robustness and reliability of peak detection.

Scientific Applications:

• Transcription factor and histone profiling: Genome-wide identification of transcription factor binding sites and histone modifications from ChIP-seq data.
• Bias characterization and correction: Modeling and correction of background biases related to mappability and GC content in ChIP-seq experiments.
• One-sample and two-sample analyses: Peak detection in single-sample experiments and comparative two-sample analyses.
• Background distribution studies: Use of naked DNA sequencing to study factors affecting background signal in ChIP-seq.

Methodology:

Implements the MOSAiCS-HMM statistical framework with a flexible mixture model for peak detection, incorporates a background model accounting for mappability and GC content, and uses naked DNA sequencing data to fit and evaluate background distributions in one-sample and two-sample analyses.