A Mis-Assembly detection and Structural Variant calling pipeline.

First it is required to pre-install conda with either anaconda or miniconda3.

anaconda - please follow these instructions

Alternatively, you can install:

miniconda3 - please follow these instructions

After installing conda, download and install the pipeline

# First clone the MASV repository
$ git clone https://github.com/Dfupa/MASV-pipeline.git
# Change to MASV directory
$ cd MASV_pipeline/
# Set the conda environment with all the necessary dependencies
$ conda env create -f MASV_pipeline.yml
# Activate the conda environment
$ conda activate MASV_pipeline

This pipline uses as an input a config file located in the MASV_pipeline/lib/config/ directory. The user can find in that directoy a script called MASV_get_config.py which will build a config file with the parameters provided by the user.

More information is provided by using the help parameter of MASV_get_config.py

# Always assuming the MASV_pipeline environment is active
$ python3 lib/config/MASV_get_config.py -h

The target rules currently available to use are:

• rule all: Outputs both Svim and Sniffles filtered calls, as well alignment QC and evaluation output (only if a truth dataset has been provided to the config file).
• rule mapping_only: Outputs only the BAM alignment between the reads and the reference.
• rule sniffles: Outputs Sniffles filtered calls.
• rule svim: Outputs Svim filtered calls.
• rule eval_sniffles: Outputs evaluation output for Sniffles only.
• rule eval_svim: Outputs evaluation output for Sniffles only.
• rule sanity_check: Outputs alignment QC based on mosdepth and Nanoplot.

In order to run the MASV pipeline just use the following command

# Always assuming the MASV_pipeline environment is active
$ snakemake -s MASV_pipeline.smk -r all -j 16 -n

The -n option is for a dry run, avoding the execution of any job yet. -j is the number of threads provided to the pipeline. In this specific example, we selected the rule all using the parameter -r.

More information regarding Snakemake and its commands can be found through Snakemake documentation.

• Right now the pipeline is able to handle .fastq and .fastq.gz. Please keep this in mind when providing your own reads.
• Right now the pipeline uses either Minimap2 or Ngmlrn aligner, but not both.
• The evaluation script requires one specific SV feature ( default is DEL) and a truth dataset for tha feature types. The user must manually edit the bedtools_eval.smk file in lib/rules/ and account for the feature change.
• The pipeline was developed and test in Unix like environments. As such, functionality is not assured when used from Windows or Mac.