hic_hiker v1.0.0

3D-DNA scaffolds refinement

Installation

$ pip install -e .

This will install the python package (and its cli command) hic_hiker.

(In the future, $ pip install hic_hiker)

Requirements

To run HiC-Hiker, you will need to install the following:

• 3D-DNA, juicer and their requirements
• python3, >=3.5 with
  • numpy, matplotlib, sklearn, scipy (basic matrix operation and visualization)
  • pandas, feather (handling of large datasets)
  • BioPython (fasta parsing)
  • pysam (sam parsing)
  • tqdm (progress bar)
  • matplotlib-scalebar (to show scalebar on the benchmark matrix plot)

(These will automatically satisfied while installation using pip)

System Requirements

• RAM >=80GB for human dataset (we tested on 100GB RAM machine)

Usage

Required Files

You have to prepare

• Input files of 3D-DNA:
  • contigs: .fasta
  • Hi-C contacts: .mnd.txt (generated by Juicer)
• Output files of 3D-DNA:
  • scaffold layout: .final.assembly You should use not .FINAL.assembly but .final.assembly
  • (chopped contigs: .final.fasta)
• An empty directory for workspace (to store intermediate files or results)

Additionally, to run benchmarks with the reference sequence (and to plot the figures on our paper), you will need:

• An Alignment file of contigs to the reference: .sam

by running:

$ bwa mem -t 32 ../hg38/hg38.fa GSE95797_Hs1.final.fasta > GSE95797_Hs1.final.fasta.sam

where GSE95797_Hs1.final.fasta is one of the outputs of 3D-DNA. (chopped contigs)

Pipeline

To refine scaffolds,

$ hic_hiker <contigs.fasta> <scaffold_layout.assembly> <contacts.mnd.txt> <workspace directory> -K <K>

where K is a threshold parameter. The Hi-C contacts whose separation distance is above K bp will ignored when polishing. For human genome dataset, we used 75000 bp.

If you want to run whole pipeline including benchmarks,

$ hic_hiker <contigs.fasta> <scaffold_layout.assembly> <contacts.mnd.txt> <workspace directory> -K <K> --refsam <final.fasta.sam>

After the process finishes, you will see in workspace directory:

• polished.assembly assembly file with refinement of orientations
• polished.fasta polished chromosome-length scaffold sequences (with no gap added)
• Figures on our paper
  • fig_distribution.png
  • fig_errorchart.png (only when you enabled benchmarks)
  • fig_matrix.png (only when you enabled benchmarks)

Uninstallation

$ pip uninstall hic_hiker

Citation

HiC-Hiker: A probabilistic model to determine contig orientation in chromosome-length scaffolds by Hi-C (not published)