HmnFusion

Getting Started

Installing

The software can be installed with conda (best way), pip and docker.
The commands worfklow-hmnfusion and workflow-fusion must be launched with the docker image hmnfusion.
To construct the BAM files, the command workflow-align must be launched with the docker image hmnfusion-align.

Install with conda

conda install -c bioconda hmnfusion

Install with pip

# Replace the <version> for the version considered
# Download from github
wget https://github.com/guillaume-gricourt/HmnFusion/releases/download/<version>/pip.zip
unzip pip.zip
# Install
pip install hmnfusion-<version>-py3-none-any.whl
# Clean up
rm pip.zip hmnfusion-<version>-py3-none-any.whl hmnfusion-<version>.tar.gz

Install with docker
Specify <version> with digits, not using latest

# Pull the image hmnfusion
docker pull ghcr.io/guillaume-gricourt/hmnfusion:<version>
or
docker pull ggricourt/hmnfusion:<version>
# Pull the image hmnfusion-align
docker pull ggricourt/hmnfusion-align:<version>

Running

Software is available by

hmnfusion command options

Commands

Extract Fusion

Aggregate results from Genefuse and Lumpy to produce a Json file.

hmnfusion extractfusion \
    --input-genefuse-json file | --input-genefuse-html file\
    --input-lumpy-vcf file \
    --output-hmnfusion-json file

Quantification

Calculate allelic frequency given postion or Json file produced by extractfusion command. A fusion is defined by two breakpoints. Only one must be in bed intervals, allelic depth is computed only on this side. Name sample is used in vcf file.

hmnfusion quantification \
    --input-hmnfusion-json file | --region chromosome:position \
    --input-sample-bam file | --input-sam file \
    --input-hmnfusion-bed file \
    --name sample_name \
    --output-hmnfusion-vcf file

MMEJ - Deletion

Extract MMEJ information from VCF file coming from classic variant caller (GATK, Varscan, …)

hmnfusion mmej-deletion \
    --input-sample-vcf file file ... \
    --input-reference-fasta file \
    --output-hmnfusion-xlsx file

MMEJ - Fusion

Extract MMEJ information from fusion breakpoints.

hmnfusion mmej-fusion \
    --input-hmnfusion-json file \
    --input-sample-bam file \
    --input-reference-fasta file \
    --name sample_name \
    --output-hmnfusion-xlsx file

Docker

Run ExtractFusion

Run extractfusion with docker like:

docker run -it \
    --rm \
    hmnfusion:latest \
    extractfusion \
    --input-genefuse-json file | --input-genefuse-html file \
    --input-lumpy-vcf file \
    --output-hmnfusion-json file

Run Workflow HmnFusion

Running combined extractfusion and quantification with one command-line:

docker run -it \
    --rm \
    hmnfusion:latest \
    workflow-hmnfusion \
    --input-genefuse-html file \
    --input-lumpy-vcf file \
    --input-sample-bam file \
    --input-hmnfusion-bed file \
    --name sample_name \
    --output-hmnfusion-vcf file

Run Workflow Fusion

Run with one command-line GeneFuse, Lumpy and HmnFusion:

docker run -it \
    --rm \
    hmnfusion:latest \
    workflow-fusion \
    --input-forward-fastq file \
    --input-reverse-fastq file \
    --input-sample-bam file \
    --output-hmnfusion-vcf file \
    --output-genefuse-html file \
    --output-lumpy-vcf file \
    --input-reference-fasta file \
    --name sample_name \
    --threads 4

Build BAM files

Only the docker image hmnfusion-align could be use for this feature.
Be aware, the size of the image is near to 15Gb.
The reference files use to build BAM files could be cited with this DOI:

docker run -it \
    --rm \
    hmnfusion-align:latest \
    workflow-align \
    --input-forward-fastq file \
    --input-reverse-fastq file \
    --output-directory file \
    --threads 4