A Method for chromatin domain partitioning based on hypergraph clustering

Figure 1. Workflow of the TORNADOES method. (a) Workflow of analyzing the chromatin domains. (b) Workflow of generating the clusters based on hypergraph learning.

datasets: Hi-C and ChIP-seq data.

Cell line Data type Data linker

GM12878 :

Hi-C https://data.4dnucleome.org/files-processed/4DNFI1UEG1HD/

H3K27ac ChIP-seq https://www.encodeproject.org/files/ENCFF087YCU/@@download/ENCFF087YCU.bigWig https://www.encodeproject.org/files/ENCFF367KIF/@@download/ENCFF367KIF.bed.gz

H3K4me3 ChIP-seq https://www.encodeproject.org/files/ENCFF480KNX/@@download/ENCFF480KNX.bigWig https://www.encodeproject.org/files/ENCFF188SZS/@@download/ENCFF188SZS.bed.gz

H3K9me3 ChIP-seq https://www.encodeproject.org/files/ENCFF533NIQ/@@download/ENCFF533NIQ.bigWig https://www.encodeproject.org/files/ENCFF682WIQ/@@download/ENCFF682WIQ.bed.gz

CTCF ChIP-seq https://www.encodeproject.org/files/ENCFF680XUD/@@download/ENCFF680XUD.bigWig https://www.encodeproject.org/files/ENCFF827JRI/@@download/ENCFF827JRI.bed.gz

POLR2A ChIP-seq https://www.encodeproject.org/files/ENCFF532WJA/@@download/ENCFF532WJA.bigWig https://www.encodeproject.org/files/ENCFF886PSD/@@download/ENCFF886PSD.bed.gz

RAD21 ChIP-seq https://www.encodeproject.org/files/ENCFF940IMA/@@download/ENCFF940IMA.bed.gz

SMC3 ChIP-seq https://www.encodeproject.org/files/ENCFF837YJA/@@download/ENCFF837YJA.bed.gz

EZH2 ChIP-seq https://www.encodeproject.org/files/ENCFF339YTO/@@download/ENCFF339YTO.bed.gz

H3K36me3 https://www.encodeproject.org/files/ENCFF268HMO/@@download/ENCFF268HMO.bed.gz

H3K4me1 https://www.encodeproject.org/files/ENCFF321BVG/@@download/ENCFF321BVG.bed.gz

H1-hESC:

Hi-C https://data.4dnucleome.org/files-processed/4DNFI2TK7L2F/

H3K27ac ChIP-seq https://www.encodeproject.org/files/ENCFF390JIZ/@@download/ENCFF390JIZ.bigWig https://www.encodeproject.org/files/ENCFF045CUG/@@download/ENCFF045CUG.bed.gz

H3K4me3 ChIP-seq https://www.encodeproject.org/files/ENCFF301RJN/@@download/ENCFF301RJN.bigWig https://www.encodeproject.org/files/ENCFF277AOQ/@@download/ENCFF277AOQ.bed.gz

H3K9me3 ChIP-seq https://www.encodeproject.org/files/ENCFF562HZQ/@@download/ENCFF562HZQ.bigWig https://www.encodeproject.org/files/ENCFF918VFL/@@download/ENCFF918VFL.bed.gz

CTCF ChIP-seq https://www.encodeproject.org/files/ENCFF147GRN/@@download/ENCFF147GRN.bigWig https://www.encodeproject.org/files/ENCFF821AQO/@@download/ENCFF821AQO.bed.gz

POLR2A ChIP-seq https://www.encodeproject.org/files/ENCFF942TZX/@@download/ENCFF942TZX.bigWig https://www.encodeproject.org/files/ENCFF322DAE/@@download/ENCFF322DAE.bed.gz

RAD21 ChIP-seq https://www.encodeproject.org/files/ENCFF883FUW/@@download/ENCFF883FUW.bed.gz

H3K36me3 https://www.encodeproject.org/files/ENCFF504KOV/@@download/ENCFF504KOV.bed.gz

H3K4me1 https://www.encodeproject.org/files/ENCFF006PXB/@@download/ENCFF006PXB.bed.gz

K562 :

Hi-C https://4dn-open-data-public.s3.amazonaws.com/fourfront-webprod/wfoutput/dcfcb009-f006-4ab8-a4c7-af72be58c12c/4DNFITUOMFUQ.hic

H3K27ac ChIP-seq https://www.encodeproject.org/files/ENCFF094XCU/@@download/ENCFF094XCU.bigWig https://www.encodeproject.org/files/ENCFF038DDS/@@download/ENCFF038DDS.bed.gz

H3K4me3 ChIP-seq https://www.encodeproject.org/files/ENCFF845BFY/@@download/ENCFF845BFY.bigWig https://www.encodeproject.org/files/ENCFF909PMV/@@download/ENCFF909PMV.bed.gz

H3K9me3 ChIP-seq https://www.encodeproject.org/files/ENCFF559MMQ/@@download/ENCFF559MMQ.bigWig https://www.encodeproject.org/files/ENCFF371GMJ/@@download/ENCFF371GMJ.bed.gz

CTCF ChIP-seq https://www.encodeproject.org/files/ENCFF268ZPN/@@download/ENCFF268ZPN.bigWig https://www.encodeproject.org/files/ENCFF221SKA/@@download/ENCFF221SKA.bed.gz

POLR2A ChIP-seq https://www.encodeproject.org/files/ENCFF914WIS/@@download/ENCFF914WIS.bigWig https://www.encodeproject.org/files/ENCFF355MNE/@@download/ENCFF355MNE.bed.gz

RAD21 ChIP-seq https://www.encodeproject.org/files/ENCFF930WPG/@@download/ENCFF930WPG.bed.gz

SMC3 ChIP-seq https://www.encodeproject.org/files/ENCFF289LLT/@@download/ENCFF289LLT.bed.gz

EZH2 https://www.encodeproject.org/files/ENCFF804RVA/@@download/ENCFF804RVA.bed.gz

H3K36me3 https://www.encodeproject.org/files/ENCFF053DAC/@@download/ENCFF053DAC.bed.gz

H3K4me1 https://www.encodeproject.org/files/ENCFF759NWD/@@download/ENCFF759NWD.bed.gz

IMR90:

Hi-C https://data.4dnucleome.org/files-processed/4DNFIH7TH4MF/

H3K27ac ChIP-seq https://www.encodeproject.org/files/ENCFF907GKJ/@@download/ENCFF907GKJ.bigWig https://www.encodeproject.org/files/ENCFF730BVO/@@download/ENCFF730BVO.bed.gz

H3K4me3 ChIP-seq https://www.encodeproject.org/files/ENCFF811ZFE/@@download/ENCFF811ZFE.bigWig https://www.encodeproject.org/files/ENCFF018CAH/@@download/ENCFF018CAH.bed.gz

H3K9me3 ChIP-seq https://www.encodeproject.org/files/ENCFF733CJA/@@download/ENCFF733CJA.bigWig https://www.encodeproject.org/files/ENCFF098XMT/@@download/ENCFF098XMT.bed.gz

CTCF ChIP-seq https://www.encodeproject.org/files/ENCFF895EDK/@@download/ENCFF895EDK.bigWig https://www.encodeproject.org/files/ENCFF307XFM/@@download/ENCFF307XFM.bed.gz

POLR2A ChIP-seq https://www.encodeproject.org/files/ENCFF444OMD/@@download/ENCFF444OMD.bigWig https://www.encodeproject.org/files/ENCFF448ZOJ/@@download/ENCFF448ZOJ.bed.gz

H3K36me3 https://www.encodeproject.org/files/ENCFF449ADN/@@download/ENCFF449ADN.bed.gz

H3K4me1 https://www.encodeproject.org/files/ENCFF611UWF/@@download/ENCFF611UWF.bed.gz

HepG2:

Hi-C https://4dn-open-data-public.s3.amazonaws.com/fourfront-webprod/wfoutput/25104375-a588-46e6-a382-663cee6c332f/4DNFICSTCJQZ.hic

H3K27ac ChIP-seq https://www.encodeproject.org/files/ENCFF847MVG/@@download/ENCFF847MVG.bigWig https://www.encodeproject.org/files/ENCFF886SZT/@@download/ENCFF886SZT.bed.gz

H3K4me3 ChIP-seq https://www.encodeproject.org/files/ENCFF359LQU/@@download/ENCFF359LQU.bigWig https://www.encodeproject.org/files/ENCFF185WOB/@@download/ENCFF185WOB.bed.gz

H3K9me3 ChIP-seq https://www.encodeproject.org/files/ENCFF284AIG/@@download/ENCFF284AIG.bigWig https://www.encodeproject.org/files/ENCFF353MWR/@@download/ENCFF353MWR.bed.gz

CTCF ChIP-seq https://www.encodeproject.org/files/ENCFF105VBF/@@download/ENCFF105VBF.bigWig https://www.encodeproject.org/files/ENCFF612ZUY/@@download/ENCFF612ZUY.bed.gz

POLR2A ChIP-seq https://www.encodeproject.org/files/ENCFF425QWO/@@download/ENCFF425QWO.bigWig https://www.encodeproject.org/files/ENCFF354VWZ/@@download/ENCFF354VWZ.bed.gz

RAD21 ChIP-seq https://www.encodeproject.org/files/ENCFF304NRB/@@download/ENCFF304NRB.bed.gz

SMC3 ChIP-seq https://www.encodeproject.org/files/ENCFF028OBL/@@download/ENCFF028OBL.bed.gz

H3K36me3 https://www.encodeproject.org/files/ENCFF489ZNJ/@@download/ENCFF489ZNJ.bed.gz

H3K4me1 https://www.encodeproject.org/files/ENCFF532AGT/@@download/ENCFF532AGT.bed.gz

Main steps:

1. Download the data set

2. Extract Hi-C data

3. Identify 2 types of TAD at different degrees

4. Identify sub-compartment in calder

5. Identify AB area with fanc tool

6. Process AB data

7. process and generate the result

8. The proportion of different chip-seq peaks anchored by different classes

9. Different types of chip-seq signal values

all experiment data and code

all experiment data and code can be downloaded from http://mged.nmdms.ustb.edu.cn/storage/data/28658183.

step1.The Hi-C matrix with resolution of 1Mb, 50kb and 25kb was extracted by juicer tool.

The IMR90 data sets for example, hic format file to download from https://data.4dnucleome.org/files-processed/4DNFIH7TH4MF/. generate tuple format: bash example/0-generateKR_hic_matrix.sh

convert tuple format into N*N matrxi format: run scripts of example/1-tuple2matrix.ipynb

step2.use the method CASPIAN to generate TAD files with different resolutions. CASPIAN code can be accessed from https://gitee.com/ghaiyan/caspian.

or you can run the code run scripts of example/2-caspian.ipynb

step3.generate the A/B compartments using fanc tool.

you can use fanc tool by accessing https://vaquerizaslab.github.io/fanc/api/analyse/compartments.html

step4.generate the sub-compartments using CALDER tool.

you can use CALDER tool by accessing https://github.com/CSOgroup/CALDER to get the chrX_sub_compartments.bed using Hi-C file with (start_location, end_location, contact_value).

library(CALDER)

for (chr in 1:22) {

chr_filename <- paste0("/home/rstudio/calder/gm12878/chr", chr, "_vc_50kb.hic")

out_dir <- paste0("./gm12878/chr", chr)

CALDER_main(chr_filename, chr=chr, bin_size=50E3, out_dir=out_dir, sub_domains=TRUE, save_intermediate_data=FALSE)

}

step5: process the epi files.

run analysis_epi.py.

step6.generate or process the hypergraph and A/B compartments, sub-compartments.

run scripts of example/3-data_process.ipynb

step7.Hypergraph learning was used to generate sub-compartments, and the similarity was compared with fanc and CALDER A/B compartment and su-compartment.

run scripts of example/4-hypergraph.ipynb

Or you can change the data of code "example/3-dataprocess.py", and run "python example/3-dataprocess.py"

step8. plot figures

run scripts of Landmark position.ipynb

step9 evaluete

run scripts of "example/5-evalute-anchor-factors.py" and "example/6-evalute-factors-density.py"