Hi-C EMT

Hi-C Extraction and Manipulation Toolkit

Tools for extracting data from .hic files to build new compact files that use the latest file version. Files can be subsampled. Regions from different files may also be stitched together.

• Note: because files will be built to use the latest v9 .hic format, you will need to make sure whatever downstream tool you are using is capable of reading v9 .hic files. The latest Juicer Tools jar, Straw, and Juicebox Jar (under Github Releases) are capable of doing so. If you are unsure whether your downstream software can read v9 files, feel free to create a Github issue to ask in this repo.

Excise

Usage

excise [-r resolution] [-c chromosomes] [--seed random_seed] [--subsample num_contacts] 
       [--cleanup] [--only-intra] <file> <out_folder>

The required arguments are:

• <file> URL or local path to .hic file from which data is to be extracted.
• <out_folder> Local folder in which to save temporary and final files.

The optional arguments are:

• -r <int> specifies highest resolution at which data will be extracted. Default: 1000.
• -c <String(s)> specifies chromosome(s) which will be extracted. Default: all chromosomes.
• --seed <long> fixes random seed for PRNG in random subsampling. Default: 0.
• --subsample <long> number of Hi-C contacts to approximately retain when subsampling file. Default: no subsampling.
• --cleanup delete temporary files (e.g. merged_no_dups) at the end. Default: keep all files.
• --only-intra only extract intra-chromosomal contacts. Default: False (i.e. extract inter and intra contacts)

Example

To subsample a map with a depth of ~5 million Hi-C contacts which goes down to 25kB resolution from GM12878_30.hic, use:

java -Xmx5g -jar hic_emt.jar excise -r 25000 --subsample 5000000 /Desktop/files/GM12878_30.hic gm_file_5M

To only subsample the first 3 chromosomes at this approximate depth (i.e. ~5 million contacts genomewide but ~600,000 contacts when only filtering for first 3 chromosomes.):

java -Xmx5g -jar hic_emt.jar excise -r 25000 -c 1,2,3 --subsample 5000000 /Desktop/files/GM12878_30.hic gm_file_5M

Stitch

Usage

stitch [-r resolution] [-k NONE/VC/VC_SQRT/KR/SCALE] [--reset-origin] [--cleanup]
       <file1,file2,...> <name1,name2,...> <chr1:x1:y1,chr2:x2:y2,...> <out_folder>

The required arguments are:

• <file1,file2,...> comma-separated list of .hic files from which data is to be extracted.
• <name1,name2,...> comma-separated list of short names/stems corresponding to the .hic files.
• <chr1:x1:y1,chr2:x2:y2,...> comma-separated list of regions from which to extract data.
• <out_folder> Local folder in which to save temporary and final files.

The optional arguments are:

• -r <int> specifies highest resolution at which data will be extracted. Default: 1000.
• [-k NONE/VC/VC_SQRT/KR/SCALE] specific normalization type to extract for the data. Default: SCALE.
• [--reset-origin] set the origin of each region at its relative start, instead of absolute coordinates. Default: use absolute coordinates.
• [--cleanup] delete temporary files (e.g. merged_no_dups) at the end. Default: keep all files.

Example

To grab a subset of KR normalized reads from chromosomes 1, 2, and 3 from three .hic files and put them in one file:

java -Xmx5g -jar hic_emt.jar stitch -r 25000 -k KR GM12878.hic,K562.hic,Hap1.hic GM,K562,Hap1 
                 1:100000000:110005000,2:115000000:125000000,3:80010000:90005000 results

Info

Usage

info <hic_file>

The required arguments are:

• <hic_file> URL or local path to .hic file for which information will be printed.

Example

To validate and print all the general information (genome, chromosomes, resolutions, normalizations, etc) about a specific .hic file:

java -Xmx5g -jar hic_emt.jar info /Desktop/files/GM12878_30.hic