ChromGene annotations for IHEC data 


by Artur Jaroszewicz and Jason Ernst (jason.ernst@ucla.edu)


ChromGene annotations for 1698 IHEC reference epigenomes based on ChIP-seq data for H3K4me3, H3K4me1, H3K27ac, H3K27me3, H3K36me3, and H3K9me3. For details about the ChromGene method, see (Jaroszewicz and Ernst, Genome Biology 2023).


Here, each gene was assigned to one of 12-mixture components, where each mixture component has two states.


The mixture components are:
1. strong_trans_enh
2. trans
3. trans_k36me3
4. trans_enh
5. trans_wk
6. ZNF
7. poised
8. PC_repr_wk
9. bivalent
10. quiescent
11. het
12. PC_repr


Descriptions of abbreviations:
* trans: Transcribed
* enh: Enhancer
* k36me3: H3K36me3
* PC_repr: Polycomb repressed
* wk: weak


Mixtures components were ordered based on decreasing order of median expression level, and states 
in mixture components were ordered by decreasing enrichment at TSS


Descriptions of mixture components:
1. strong_trans_enh: Strongly transcribed genes with enhancer marks H3K4me1 and H3K27ac, highest expressed
2. trans: Transcribed genes, highly expressed. Lower levels of H3K36me3 and H3K4me1 than strong_trans_enh
3. trans_k36me3: Transcribed genes, moderately expressed, marked with H3K36me3. Relatively low levels of enhancer marks
4. trans_enh: Transcribed genes, moderately expressed, marked with enhancer marks H3K4me1 and to a lesser extent H3K27ac. Relatively low levels of H3K36me3
5. trans_wk: Transcribed genes, lowly expressed, marked with enhancer mark H3K27ac and to a lower extent, H3K4me3. Relatively low levels of H3K36me3
6. ZNF: Zinc Finger-related genes, with a state marked by H3K36me3 and H3K9me3
7. poised: lowly-expressed genes with high levels of H3K4me1, but low levels of H3K36me3
8. PC_repr_wk: lowly expressed genes with high levels of H3K4me1, weakly marked with polycomb repressive mark H3K27me3
9. bivalent: lowly expressed genes, marked with high level of repressive mark H3K27me3, and moderate levels of H3K4me1 and H3K4me3
10. quiescent: lowly expressed genes, lack of significant histone marks
11. het: generally unexpressed genes associated with heterochromatin
12. PC_repr: generally unexpressed genes, marked with polycomb repressive mark H3K27me3


Files:
* IHEC_ChromGene_assts.named.tsv.gz - Tab-delimited file containing assignments of protein-coding genes to ChromGene mixture components. The first column is the Ensembl gene ID, and the next seven columns contain information about the genes in bed format. Gene annotations were based on Gencode v29. The remaining columns correspond to ChromGene assignments to mixture components, one column for each epigenome. 

* IHEC_ChromGene_assts.numbered.tsv.gz - The same file as IHEC_ChromGene_assts.named.tsv.gz, except using mixture components numbers instead of mixture component names. 

* emissions_25.txt, emissions_25.png, emissions_25.svg - Shows the emission parameters of the model in txt, png, and svg formats, respectively.

* transitions_25.txt, transitions_25.png, transitions_25.svg - Shows the transitions parameters of the model in txt, png, and svg formats respectively.

* model_25.txt  - text file of all the model parameters in a format that can be run with ChromGene.