HiNT

A computational method for detecting copy number variations and translocations from Hi-C data

Summary

HiNT (Hi-C for copy Number variation and Translocation detection), a computational method to detect CNVs and Translocations from Hi-C data. HiNT has three main components: HiNT-PRE, HiNT-CNV, and HiNT-TL. HiNT-PRE preprocesses Hi-C data and computes the contact matrix, which stores contact frequencies between any two genomic loci; both HiNT-CNV and HiNT-TL starts with HI-C contact matrix, predicts copy number segments, and inter-chromosomal translocations, respectively

Overview of HiNT workflow:

Installation

Dependencies

R and R packages

1. R >= 3.4
2. mgcv, strucchange, doParallel, Cairo, foreach

Python and Python packages

1. python >= 3.5
2. pyparix >= 0.3.0, cooler >= 0.7.4, pairtools >= 0.2.2, numpy, scipy, pandas, sklearn, multiprocessing

Java and related tools (Optional: required when want to process Hi-C data with juicer tools)

1. Java (version >= 1.7)
2. Juicer tools (1.8.9 is recommended)

Perl

1. Perl (version >= 5)

Other dependencies

1. samtools (1.3.1+)
2. BIC-seq2 (0.7.3) ! This is optional: if you don't want to run HiNT-CNV, you don't need this package. [Download BICseq2, unzip it, and give the path of BICseq2-seg_v0.7.3 (/path/to/BICseq2-seg_v0.7.3)].
3. bwa (0.7.16+) ! This is optional: required only when your input is fastq
4. tabix (0.2.6)

Install HiNT

• Method1: Install using conda (highly recommended)

  $ conda install -c su hint

  or

  $ conda install hint

• Method2: Install from PyPI using pip.

  $ pip install HiNT-Packages

• Method3: Install manually

  1. Install HiNT dependencies
  2. Download HiNT git clone https://github.com/parklab/HiNT.git
  3. Go to HiNT directory, install it by $ python setup.py install

  *** Type $ hint to test if HiNT successfully installed

• Method 4: Run HiNT in a Docker container (highly recommended)

  $ docker pull suwangbio/hint

  $ docker run suwangbio/hint hint

  See details of the usage on HiNT page at docker hub

Download reference files used in HiNT HERE

1. Download HiNT references HERE. Only hg19, hg38 and mm10 are available currently. Unzip it $ unzip hg19.zip
2. Download HiNT background matrices HERE. Only hg19, hg38 and mm10 are available currently. Unzip it $ unzip hg19.zip
3. Download BWA index files HERE. Only hg19, hg38 and mm10 are available currently. Unzip it $ unzip hg19.zip

Quick Start

• Download the test datasets from HERE

HiNT-PRE

HiNT pre: Preprocessing Hi-C data. HiNT pre does alignment, contact matrix creation and normalization in one command line.

$ hint pre -d /path/to/hic_1.fastq.gz,/path/to/hic_2.fastq.gz -i /path/to/bwaIndex/hg19/hg19.fa --refdir /path/to/refData/hg19 --informat fastq --outformat cooler -g hg19 -n test -o /path/to/outputdir --pairtoolspath /path/to/pairtools --samtoolspath /path/to/samtools --coolerpath /path/to/cooler

$ hint pre -d /path/to/test.bam --refdir /path/to/refData/hg19 --informat bam --outformat juicer -g hg19 -n test -o /path/to/outputdir --pairtoolspath /path/to/pairtools --samtoolspath /path/to/samtools --juicerpath /path/to/juicer_tools.1.8.9_jcuda.0.8.jar

use $ which samtools $ which pairtools $ which cooler to get the absolute path of these tools, and /path/to/juicer_tools.1.8.9_jcuda.0.8.jar should be the path where you store this file

see details and more options

$ hint pre -h

HiNT-CNV

HiNT cnv: prediction of copy number information, as well as segmentation from Hi-C.

$ hint cnv -m contactMatrix.cool -f cooler --refdir /path/to/refDir/hg19 -r 50 -g hg19 -n test -o /path/to/outputDir --bicseq /path/to/BICseq2-seg_v0.7.3 -e MboI

$ hint cnv -m /path/to/4DNFIS6HAUPP.mcool::/resolutions/50000 -f cooler --refdir /path/to/refDir/hg38 -r 50 -g hg38 -n HepG2 --bicseq /path/to/BICseq2-seg_v0.7.3 -e DpnII --maptrack 36mer

$ hint cnv -m /path/to/4DNFICSTCJQZ.hic -f juicer --refdir /path/to/refDir/hg38 -r 50 -g hg38 -n HepG2 --bicseq /path/to/BICseq2-seg_v0.7.3 -e DpnII

$ hint cnv -m /path/to/4DNFICSTCJQZ.hic -f juicer --refdir /path/to/refDir/hg38 -r 50 -g hg38 -n HepG2 --bicseq /path/to/BICseq2-seg_v0.7.3 -e DpnII --doiter

/path/to/BICseq2-seg_v0.7.3 should be the path where you store this package

see details and more options

$ hint cnv -h

HiNT-TL

HiNT tl: interchromosomal translocations and breakpoints detection from Hi-C inter-chromosomal interaction matrices.

$ hint tl -m /path/to/data_1Mb.cool,/path/to/data_100kb.cool --chimeric /path/to/test_chimeric.sorted.pairsam.gz --refdir /path/to/refDir/hg19 --backdir /path/to/backgroundMatrices/hg19 --ppath /path/to/pairix -f cooler -g hg19 -n test -o /path/to/outputDir

$ hint tl -m /path/to/4DNFIS6HAUPP.mcool::/resolutions/1000000,/path/to/4DNFIS6HAUPP.mcool::/resolutions/100000 -f cooler --refdir /path/to/refDir/hg38 --backdir /path/to/backgroundMatrices/hg38 -g hg38 -n 4DNFICSTCJQZ -c 0.05 --ppath /path/to/pairix -p 12

$ hint tl -m /path/to/4DNFICSTCJQZ.hic -f juicer --refdir /path/to/refData/hg38 --backdir /path/to/backgroundMatrices/hg38 -g hg38 -n 4DNFICSTCJQZ -c 0.05 --ppath /path/to/pairix -p 12 -o HiNTtransl_juicerOUTPUT

use $ which pairix to get the absolute path of pairix

see details and more options

$ hint tl -h

Output of HiNT

HiNT-PRE output

In the HiNT-PRE output directory, you will find

1. jobname.bam aligned lossless file in bam format
2. jobname_merged_valid.pairs.gz reads pairs in pair format
3. jobname_chimeric.sorted.pairsam.gz ambiguous chimeric read pairs used for breakpoint detection in pairsam format
4. jobname_valid.sorted.deduped.pairsam.gz valid read pairs used for Hi-C contact matrix creation in pairsam format
5. jobname.mcool Hi-C contact matrix in cool format
6. jobname.hic Hi-C contact matrix in hic format

HiNT-CNV output

In the HiNT-CNV output directory, you will find

1. jobname_GAMPoisson.pdf the GAM regression result
2. segmentation/jobname_bicsq_allchroms.txt CNV segments with log2 copy ratio and p-values in txt file
3. segmentation/jobname_resolution_CNV_segments.png figure to visualize CNV segments
4. segmentation/jobname_bicseq_allchroms.l2r.pdf figure to visualize log2 copy ration in each bin (bin size = resolution you set)
5. segmentation/other_files intermediate files used to run BIC-seq
6. jonname_dataForRegression/* data used for regression as well as residuals after removing Hi-C biases

HiNT-TL output

In the HiNT-TL output directory, you will find

1. jobname_Translocation_IntegratedBP.txt the final integrated translocation breakpoint
2. jobname_chrompairs_rankProduct.txt rank product predicted potential translocated chromosome pairs
3. otherFolders intermediate files used to identify the translocation breakpoints