HiC-pipeline user guide

The pipeline takes a list of HiC cram/bam and a reference fasta file (the fasta file should be named something.fa) as input files and produce pretext file and mcool file to upload to higlass server.

software and dependencies

(1) Install conda and snakemake

conda activate snake_env
conda install -c bioconda snakemake

(2) software

• vr-runner download the vr-runner from github

https://github.com/VertebrateResequencing/vr-runner

Modify you PERL5LIB

export PERL5LIB="$HOME/git/vr-runner/modules:$PERL5LIB"

• samtools
• pretext
• bwa
• bamToBed
• picard
• bammarkduplicates2
• cooler

https://github.com/mirnylab/cooler

usuage

The pipeline can be run under snakemake conda environment or without:

To run the pipeline under conda environment

A yaml configuration file need to be completed if you choose this option. The yaml file format is tailored to the VGP analysis pipeline in GRIT processing other analysises, such as break10x and bionano. If you only run hic analysis please only alter the HiC session in the yaml file.

To fill in sample information.

  assembly:
  level: scaffold
  sample: fAnaAna1
  class: fish
  dbVersion: "1"
  asmVersion: test
  gevalType: VGP
  reference: /lustre/scratch115/teams/grit/users/yy5/hic/fish/ref/ref.fa

Prepare a list of hic data file

hic:
    hicreads: /lustre/scratch115/dove.fofn

The dove.fofn contains the list of unmapped HiC cram/bam files should look like:

/lustre/scratch116/vr/projects/vgp/build/fish/fEleEle1/Arima/crams/fEleEle1_ARIMA241127L002.cram
/lustre/scratch116/vr/projects/vgp/build/fish/fEleEle1/Arima/crams/fEleEle1_ARIMA241127L005.cram
/lustre/scratch116/vr/projects/vgp/build/fish/fEleEle1/Arima/crams/fEleEle1_ARIMA241127L007.cram

Amend the path to the runner script (run-hic script in runners folder)

tools:
  runner_path: /nfs/team135/yy5/geval_pipe_dev/runner_scripts

You may also need to change the cluster information in the cluster.yaml.

Run the pipeline without conda environment

The basic usuage to run the run-hic script as below:

run-hic +loop 60 -f /path/to/dove.fofn -s sample_name -q 0 -r /lustre/scratch115/ref/ref.fa -o /lustre/scratch115/teams/grit/users/yy5/hic/fish/hic -z hic_done

The arguments needed are:

-s (sample_name), eg. fAnaAna1
-f (reference)
-o (output directory)
-q filter by quality score, to keep the multimapped reads, please set -q 0
-z (the name of file that indicate the procedure finishes)