ChromHMM results from Jason Ernst Lab (jason.ernst@ucla.edu) (Ha Vu and Jason Ernst)
=======================================================

A 'univeral' chromatin state annotation based on a 100-state full-stack model were generated by applying the ChromHMM (Ernst and Kellis, 2012) (v1.24) LearnModel to all 5339 observed data-sets for H3K4me1, H3K4me3, H3K9ac, H3K27ac, H3K27me3, H3K36me3 following the approach of Vu and Ernst, 2022. This provides a single annotation of the genome for each position based on all the input datasets.

In V1 of these annotations state labels, colors, and ordering were automatically generated by determining for each state of this model the state for which it has maximum enrichment based on the state annotations from the Vu and Ernst, 2022 model after lifting over the annotations from hg19 to hg38. If multiple states mapped to the same state from Vu and Ernst, 2022 letter suffixes were appended and ordered alphabetically based on decreasing enrichment values.

File IHEC_genome_100_ChromHMM_hg38_V1.bed.gz - contains the universal chromatin state annotations in a four column bed format
File IHEC_genome_100_ChromHMM_hg38_V1_browser.bed.gz - contains the universal chromatin state annotations in a bed format for viewing in a genome browser
File IHEC_genome_100_ChromHMM_hg38_V1.bb.gz - contains the universal chromatin state annotations in a BigBed format for viewing in a genome browser
File model_100_stacked_V1.txt - ChromHMM model file with the states ordered based on the V1 ordering
File emissions_100_stacked_V1.txt - emission parameter file with the states ordered basded on the V1 ordering and marks based on ChromHMM's emission parameter ordering approach
File transitions_100_stacked_V1.txt - transition parameter file with the states ordered basded on the V1 ordering