A Tutorial on Computational Methods for Modeling and Analyzing Epigenomic Data

Explore computational methods for modeling and analyzing epigenomic data in this 45-minute tutorial from the Computational Genomics Summer Institute. Delve into supervised integration of multiple chromatin markers using multivariate Hidden Markov Models, specifically focusing on the ChromHMM approach. Examine the ENCODE project's study of nine marks in nine human cell lines and discover how chromatin state annotations reveal cell type-specific enrichments of GWAS hits. Learn about regulatory activity in DNA accessible sites and its dependence on chromatin state. Compare three options for applying ChromHMM to multiple cell types, including the advantages of stacked modeling over concatenated models. Investigate the application of ChromGene to Roadmap Epigenomics data and its relationship with gene expression. Gain insights from related research papers on genome annotation, human-mouse conservation scoring, and large-scale imputation of epigenomic datasets for systematic tissue annotation.

Intro
The epigenome provides cell type specific markers
Supervised Integration of Multiple Chromatin Marks
Approach: Multivariate Hidden Markov Model (ChromHMM)
ENCODE: Study nine marks in nine human cell lines
Chromatin state annotations exhibit cell type specific enrichments of GWAS hits
Regulatory Activity in DNA Accessible Sites Depends on Chromatin State
Three Options for Applying ChromHMM to Multiple Cell Types
Stacked Modeling Provides Complementary Advantanges to Concatenated Model
ChromGene applied to Roadmap Epigenomics data
ChromGene annotations relationship with gene expression