Massively Parallel and Portable Genomic Sequence Analysis (Mappgene)

Mappgene is a genomic sequence analysis workflow designed for high-performance computing. It currently wraps V-pipe (https://github.com/cbg-ethz/V-pipe) with a collection of useful scripts for deployment in almost any Linux environment.

Quick Install

pip3 install --user parsl
git clone https://github.com/LLNL/mappgene.git

Download the Singularity image (https://drive.google.com/file/d/1qOhgiQChtfaMk0VNYpkZvVIYNiGs9ItI/view?usp=sharing) and place it at container/image.sif.

Running

Organizing your data

Copy your reference genome in fasta format to vpipe_files/references/.
Inspect and configure vpipe_files/vpipe.config for your run (e.g., reference = , trim_percent_cutoff = , threads = )
Organize your input files (typically gzip compressed fastq formatted paired-end reads). mappgene.py expects the following layout, where sample names are taken from the subdirectories (i.e., foo, bar, baz):
```
/path/to/input_dirs
|-- foo
|   |-- foo_R1.fastq.gz
|   `-- foo_R2.fastq.gz
|-- bar
|   |-- bar_R1.fastq.gz
|   `-- bar_R2.fastq.gz
|-- baz
|   |...
...
```

Run

Specify arguments

Example:

python3 mappgene.py \
    --input_dirs /path/to/input_dirs \
    --output_dirs /path/to/output_dirs \
    --read_length 130 \
    --nnodes 16 \
    --walltime 12:59:59

python3 mappgene.py <config_json>

More info

python3 mappgene.py --help

How it works

V-pipe and other software have been pre-installed in a container (https://github.com/hpcng/singularity).
User parameters are parsed from command-line or configuration JSON (e.g. configs/example/catalyst.json). Note that every parameter has a default value (see mappgene.py).
Software tasks are then distributed to compute nodes in parallel (https://github.com/Parsl/parsl).

License

Mappgene is distributed under the terms of the BSD-3 License.

LLNL-CODE-821512