Imputed signal tracks

DATA SOURCE

Download URL:
Imputed Signal Tracks for histone ChIP-seq, DNase-seq, RNA-seq and methylation

hg19 for all marks

GRCh38 lift-over for 12 marks

For imputing epigenomic data sets, we used a new method, ChromImpute (Ernst and Kellis, Nature Biotech 2015), that to predict a target mark in a target reference epigenome combines information about other marks mapped in the target reference epigenome, and the target mark at the same position in similar reference epigenomes through an ensemble of regression trees to make predictions about unobserved datasets. Additionally an imputed version of each observed data set was generated without using the corresponding observed data. Each directory corresponds to a Mark and contains a set of BigWig files one for each reference epigenome. The signal tracks for the Histone Modifications and DNase are based on the p-value signal tracks. We also provide signal tracks for GRCh38 lift-over for 12 marks (DNase, H2A.Z, H3K4me1, H3K4me2, H3K4me3, H3K9ac, H3K9me3, H3K27ac, H3K27me3, H3K36me3, H3K79me2 and H4K20me1).

Peak calls on imputed data

DATA SOURCE

Download URL:
Peak calls on imputed data

Narrow contiguous regions of enrichment (peaks) for histone ChIP-seq and DNase-seq
Data format: BED
or
Open in a new page (deactivate pop-up blockers)
Broad domains on enrichment for histone ChIP-seq
Data format: BED
or
Open in a new page (deactivate pop-up blockers)

Macs2 (version. 2.0.10) was used to call peaks on the imputed signal data. The bdgpeakcall command was used to generate narrowPeaks while the bdgbroadcall command was used to generate gappedPeaks with the '-c' cutoff flag set to 2. The peak calls are reported in a three column bed format.

Chromatin state model based on imputed data (25 state, 12 marks, 127 epigenomes)

DATA SOURCE

12-mark/127-reference epigenome/25-state Imputation Based Chromatin State Model

Summarized visualization of all 127 epigenomes using epilogos

The states are as follows

STATE NO.	MNEMONIC	DESCRIPTION	COLOR NAME	COLOR CODE
1	TssA	Active TSS	Red	255,0,0
2	PromU	Promoter Upstream TSS	Orange Red	255,69,0
3	PromD1	Promoter Downstream TSS 1	Orange Red	255,69,0
4	PromD2	Promoter Downstream TSS 2	Orange Red	255,69,0
5	Tx5	Transcribed - 5' preferential	Green	0,128,0
6	Tx	Strong transcription	Green	0,128,0
7	Tx3	Transcribed - 3' preferential	Green	0,128,0
8	TxWk	Weak transcription	Lighter Green	0,150,0
9	TxReg	Transcribed & regulatory (Prom/Enh)	Electric Lime	194,225,5
10	TxEnh5	Transcribed 5' preferential and Enh	Electric Lime	194,225,5
11	TxEnh3	Transcribed 3' preferential and Enh	Electric Lime	194,225,5
12	TxEnhW	Transcribed and Weak Enhancer	Electric Lime	194,225,5
13	EnhA1	Active Enhancer 1	Orange	255,195,77
14	EnhA2	Active Enhancer 2	Orange	255,195,77
15	EnhAF	Active Enhancer Flank	Orange	255,195,77
16	EnhW1	Weak Enhancer 1	Yellow	255,255,0
17	EnhW2	Weak Enhancer 2	Yellow	255,255,0
18	EnhAc	Primary H3K27ac possible Enhancer	Yellow	255,255,0
19	DNase	Primary DNase	Lemon	255,255,102
20	ZNF/Rpts	ZNF genes & repeats	Aquamarine	102,205,170
21	Het	Heterochromatin	Light Purple	138,145,208
22	PromP	Poised Promoter	Pink	230,184,183
23	PromBiv	Bivalent Promoter	Dark Purple	112,48,160
24	ReprPC	Repressed Polycomb	Gray	128,128,128
25	Quies	Quiescent/Low	White	255,255,255

A chromatin state model based on the imputed data for 12 marks, H3K4me1, H3K4me2, H3K4me3, H3K9ac, H3K27ac, H4K20me1, H3K79me2, H3K36me3, H3K9me3, H3K27me3, H2A.Z, and DNase, across all 127 reference epigenomes with 25-states was learned. Within the directory for this model one can find both plain bed files containing the segmentation information and versions that can be viewed in the browser. An additional chromatin state model based on 29-marks across just the seven class 1 reference epigenomes was also learned.