Skip to content

Latest commit

 

History

History
127 lines (91 loc) · 4.83 KB

Singularity.md

File metadata and controls

127 lines (91 loc) · 4.83 KB
Raw

Singularity workflow

Snakemake + Singularity is a convenient way to run a complex workflow without spending too much time on installing software. Here, we provide an image called mosaicatcher-pipeline to easily run the MosaiCatcher workflow.

Requirements

  • Snakemake version 5.3.0+ (tested 5.3.0). Version 5.3.1 even fixes some bugs that we had to circumvent (not tested).
  • Singularity. Tested with version 2.5.2 and 3.0

Note that Snakemake 5.3.0 doesn't work well with Singularity 3.0+ because of the format of the version string of Singularity. It can be circumvented by creating a wrapper script around the singularity command. This was fixed in version 5.3.1.

Singularity/Docker image

We created a Docker image mosaicatcher-pipeline containing all software tools for the Mosaicatcher workflow (except Snakemake). Snakemake can make use of it when run in --use-singularity mode. The link to this Docker image is already hardcoded inside the Snakefile.

To reduce the file size of this image (it is now ~2GB), the reference genome (GRCh38) was stripped off. Hence these files have to be provided via -B when the pipeline is run (see below).

Running Snakemake in Singularity mode

  1. (Optional) Download example data

    You can download some of the data included in our study via the enaBrowserTools:

    enaGroupGet -g read -f submitted PRJEB30027

    These are the available files:

    • RPE1 wild type line (sample name RPE1-WT): ERR2940244 (RPE1WTPE20401.sort.mdup.bam) ~ ERR2940323 (RPE1WTPE20495.sort.mdup.bam) (80 cells)
    • BM510 line (sample name RPE-BM510): ERR2940324 (BM510x04_PE20301.sort.mdup.bam) ~ ERR2940468 (BM510x3PE20496.sort.mdup.bam) (145 cells)
    • C7 line (sample name C7_data): ERR2940469 (C7x02PE20301.sort.mdup.bam) ~ ERR2940622 (C7x03PE20396.sort.mdup.bam) (154 cells)

  2. Download this pipeline

    git clone https://github.com/friendsofstrandseq/mosaicatcher-pipeline pipeline
cd pipeline

  3. Add your data

    I.e. bam files and SNV calls (if available). See instructions in the README

    We noticed a problem with soft-linked BAM files when using Singularity. Hence it is recommended to copy or hard-link BAM files into the bam/xxx/all and bam/xxx/selected folders.

  4. Adapt the config file

    As described in the README, but now use Snake.config-singularity.json which already has software paths set correctly. This is because all required software is contained in the Singularity image.

  5. Provide the reference genome

    The singularity image does not yet contain the reference genome. You will need to provide it from outside during execution.

    To do so, set two variables in your bash:

    • REF path to the refernce genome FASTA file
    • R_REF The reference genome of the R-package BSgenome.Hsapiens.UCSC.hg38. You might have to install this package first to get this file.

    For example on my system, I would type

    REF="/home/meiers/referece/GCA_000001405.15_GRCh38_no_alt_analysis_set.fna"
R_REF="/home/meiers/R/x86_64-pc-linux-gnu-library/3.5/BSgenome.Hsapiens.UCSC.hg38/extdata/single_sequences.2bit"

    These variables will be passed on to Singularity via the -B flag in the next step.

  6. Execute Snakemake

    Run Snakemake in Singularity mode similar to this (also provided in a script).

    # Please first set REF and R_REF

snakemake \
    -j 2 \
    --configfile Snake.config-singularity.json \
    --use-singularity \
    --singularity-args \
       "-B ${REF}:/reference.fa:ro \
	-B ${REF}.fai:/reference.fa.fai:ro \
	-B ${R_REF}:/usr/local/lib/R/site-library/BSgenome.Hsapiens.UCSC.hg38/extdata/single_sequences.2bit:ro" \
    --latency-wait 60 \
    --printshellcmd

Singularity vs. Docker workflow

In the Docker workflow containing an example data set, Snakemake is included within the image. In fact, the mosaicatcher-pipeline-rpe-1 image is based on the mosaicatcher-pipeline used here.

The major difference is that here, snakemake is run on your system and not from within a Docker container. This gives you easy access to all other functionality of Snakemake, including multi-core and cluster support. Only each job within the workflow will be run inside the image.