MICC

MICC detects chromatin interactions from high-resolution ChIA-PET sequencing data by applying a Bayesian mixture model to distinguish true interactions from random ligation and collision noise.

Key Features:

• Bayesian Mixture Model: Employs a Bayesian mixture model to separate true chromatin interactions from background noise caused by random ligation or collision events.
• Noise Filtering: Systematically filters noise intrinsic to ChIA-PET experiments, specifically addressing random ligation and collision artifacts.
• High Sensitivity: Identifies chromatin interactions with significantly higher sensitivity compared to existing methods while maintaining the same false discovery rate.
• R Package Implementation: Implemented and distributed as an R package for computational analysis of ChIA-PET data.

Scientific Applications:

• Mapping chromatin loops and domains: Maps chromatin loops and domains from ChIA-PET data to define long-range genomic interactions.
• Investigating gene regulatory networks: Supports analysis of gene regulatory networks mediated by long-range chromatin interactions.
• Epigenetics: Facilitates studies of long-range chromatin interactions underlying epigenetic regulation.
• Developmental biology: Enables analysis of chromatin interactions relevant to developmental gene regulation.
• Disease mechanisms (e.g., cancer): Aids in identifying interaction changes associated with disease mechanisms such as cancer.
• Structural variation studies: Supports examination of structural variation effects on chromatin architecture.

Methodology:

Applies a Bayesian mixture model to ChIA-PET sequencing data to distinguish true interactions from background arising from random ligation and collision events; implemented as an R package.