Integrative analysis of histone ChIP-seq and transcription data using Bayesian mixture models

Area of Science:

• Epigenetics
• Bioinformatics
• Genomics

Background:

• Histone modifications regulate gene transcription.
• Integrated analysis of histone modification (ChIP-seq) and gene transcription (RNA-seq/microarrays) data is challenging.
• Existing methods for integrative analysis are limited.

Purpose of the Study:

• To develop a novel bioinformatics approach for the integrative analysis of histone modification and gene transcription data.
• To identify genes with differential transcript abundances potentially caused by altered histone modifications.
• To provide a robust method for comparing gene expression and epigenetic states across conditions.

Main Methods:

• Introduced a correlation measure for integrating ChIP-seq and gene transcription data.
• Emphasized the importance of proper ChIP-seq data normalization.
• Applied Bayesian mixture models for analyzing the correlation measure distribution.
• Utilized implicit classification from mixture models to detect differential genes.
• Validated the method on diverse biological datasets.

Main Results:

• Demonstrated the effectiveness of the proposed correlation measure for integrative analysis.
• Showcased the critical role of ChIP-seq data normalization in achieving reliable results.
• Successfully identified genes exhibiting differences in both transcription and histone modification across conditions.
• Proved the superiority of the novel integrated approach compared to separate data analyses.

Conclusions:

• The developed bioinformatics approach enables robust integrative analysis of histone modification and gene transcription data.
• The method accurately detects genes affected by epigenetic changes.
• The R/Bioconductor package 'epigenomix' implements this novel approach, facilitating its application in biological research.