pcawg-muse

PCAWG MuSE variant calling workflow is developed by MD Anderson Cancer Center (http://bioinformatics.mdanderson.org/main/MuSE), it consists of software component calling structural variants using uniformly aligned tumor / normal WGS sequences. The workflow has been dockerized and packaged using CWL workflow language, the source code is available on GitHub at: https://github.com/ICGC-TCGA-PanCancer/pcawg-muse. ## Run the workflow with your own data ### Prepare compute environment and install software packages The workflow has been tested in Ubuntu 16.04 Linux environment with the following hardware and software settings. #### Hardware requirement (assuming 30X coverage whole genome sequence) - CPU core: 16 - Memory: 64GB - Disk space: 1TB #### Software installation - Docker (1.12.6): follow instructions to install Docker https://docs.docker.com/engine/installation - CWL tool ``` pip install cwltool==1.0.20170217172322 ``` ### Prepare input data #### Input aligned tumor / normal BAM files The workflow uses a pair of aligned BAM files as input, one BAM for tumor, the other for normal, both from the same donor. For improved calling result we pre-process the aligned BAMs using PCAWG GATK Co-cleaning workflow, see [here](https://dockstore.org/workflows/ICGC-TCGA-PanCancer/pcawg-gatk-cocleaning) for more information how to run it. Here we assume aligned pre-processed BAMs are *tumor_sample.bam* and *normal_sample.bam*, and are under *bams* subfolder. #### Reference data files The workflow also uses two reference files as input, both can be downloaded from the ICGC Data Portal: - reference genome sequence *genome.fa.gz* under https://dcc.icgc.org/releases/PCAWG/reference_data/pcawg-bwa-mem - known dbSNP entries *dbsnp_132_b37.leftAligned.vcf.gz* under https://dcc.icgc.org/releases/PCAWG/reference_data/pcawg-muse We assume the reference files are under *reference* subfolder. #### Job JSON file for CWL Finally, we need to prepare a JSON file with input, reference files specified. Please replace the *tumor* and *normal* parameters with your real BAM files. Name the JSON file: *pcawg-muse-variant-caller.job.json* ``` { "mode": "wgs", "tumor": { "class": "File", "location": "bams/tumor_sample.bam" }, "normal": { "class": "File", "location": "bams/normal_sample.bam" }, "reference": { "class": "File", "location": "reference/genome.fa.gz" }, "known": { "class": "File", "location": "reference/dbsnp_132_b37.leftAligned.vcf.gz" } } ``` ### Run the workflow #### Option 1: Run with CWL tool - Download CWL workflow definition file ``` wget -O pcawg-muse-variant-caller.cwl "https://raw.githubusercontent.com/ICGC-TCGA-PanCancer/pcawg-muse/0.1.2/muse.cwl" ``` - Run `cwltool` to execute the workflow ``` nohup cwltool --debug --non-strict pcawg-muse-variant-caller.cwl pcawg-muse-variant-caller.job.json > pcawg-muse-variant-caller.log 2>&1 & ``` #### Option 2: Run with the Dockstore CLI See the *Launch with* section below for details.

URL(s):

• https://dockstore.org/api/api/ga4gh/v2/tools/quay.io%2Fpancancer%2Fpcawg-mu...