A flexible and powerful bayesian hierarchical model for ChIP-Chip experiments.

Area of Science:

• Genomics
• Bioinformatics
• Statistical Genetics

Background:

• Chromatin-immunoprecipitation microarrays (ChIP-chip) are crucial for identifying DNA-binding protein regions.
• ChIP-chip data analysis faces challenges due to high probe numbers, noise, and spatial probe dependence.

Purpose of the Study:

• To develop a robust statistical method for detecting transcription factor bound regions in ChIP-chip data.
• To address the challenges of noise, spatial dependence, and probe outliers in ChIP-chip analysis.

Main Methods:

• Proposed Bayesian analysis of ChIP-chip (BAC), a method incorporating probe dependence and robust to outliers.
• Utilized Markov chain Monte Carlo for parameter estimation and posterior probabilities for bound region detection.
• Incorporated an exchangeable prior for variances to manage probe-specific variability.

Main Results:

• BAC demonstrated strong performance in detecting experimentally validated regions on two public ChIP-chip datasets.
• BAC showed superior power and robustness to model misspecification compared to Wilcoxon's rank sum test, TileMap, HGMM, and MAT.
• BAC was found to be less sensitive to probe outliers than HGMM.

Conclusions:

• BAC is a powerful and robust statistical method for analyzing ChIP-chip data.
• The method provides well-calibrated posterior probabilities for accurate false discovery rate estimation.
• BAC offers an effective solution for identifying transcription factor binding sites in genomic studies.