search

bsmn / bsmn-pipeline

BSMN common data processing pipeline
11 stars 10 forks source link

readme

DOI

bsmn_pipeline

BSMN common data processing pipeline implementing various SGE (Sun Grid Engine) jobs arranged for genome alignment, variant calling and filtering.

Setup and installation

This pipeline can be run in any cluster system using SGE job scheduler. We would recommend set your own cluster in AWS using AWS ParallelCluster.

AWS ParallelCluster

For installing and setting up parallelcluster, please see the Getting Started Guide for AWS ParallelCluster.

Extra set up for SGE

The pipeline require a parallel environment named "threaded" in your SGE system. If your SGE system doen't have this parallel environment, you should add it into yours.

$ cat >threaded.conf <<END
pe_name            threaded
slots              99999
user_lists         NONE
xuser_lists        NONE
start_proc_args    NONE
stop_proc_args     NONE
allocation_rule    \$pe_slots
control_slaves     FALSE
job_is_first_task  TRUE
urgency_slots      min
accounting_summary TRUE
qsort_args         NONE
END
$ sudo su
# qconf -Ap threaded.conf
# qconf -mattr queue pe_list threaded all.q

Installing pipeline

Clone this repository where you want it installed in your cluster. If you work with an m5.large type AWS EC2 instance we recommend the file systems mounted at /shared or /efs.

cd /shared
git clone https://github.com/bsmn/bsmn_pipeline

Create a conda environment from YAML file to install software dependencies running the following commands. By default, the name of environment will be bp. you can change it by adding a -n your_name option.

conda env create -f /path/to/pipeline/environment.yml

Instead you can use a different YAML file for a version-fixed conda environment where major tools for alignment and variant calling are frozen with their exact versions we used in BSMN data analyses as follows:

By default, the name of frozen environment will be bp_frozen.

conda env create -f /path/to/pipeline/environment_frozen.yml

Install ucsc-* software packages.

conda install -n bp -c bioconda ucsc-fetchchromsizes ucsc-bigwigaverageoverbed ucsc-wigtobigwig ucsc-liftover
# If you are on bp_frozen
conda install -n bp_frozen -c bioconda ucsc-fetchchromsizes ucsc-bigwigaverageoverbed ucsc-wigtobigwig ucsc-liftover

Due to license restrictions, you need to download a copy of GATK3 from the Broad Institute.

conda activate bp # Make sure you've activated the environment you are working on.
wget -qO- https://storage.googleapis.com/gatk-software/package-archive/gatk/GenomeAnalysisTK-3.8-1-0-gf15c1c3ef.tar.bz2 \
     |tar xj --strip=1 */GenomeAnalysisTK.jar
gatk3-register GenomeAnalysisTK.jar
rm GenomeAnalysisTK.jar # Once register, you can delete the downloaded file.

If you are on the frozen environment (bp_frozen), you should download a copy of verion 3.7.

conda activate bp_frozen # Make sure you've activated the environment you are working on.
wget -qO- https://storage.googleapis.com/gatk-software/package-archive/gatk/GenomeAnalysisTK-3.7-0-gcfedb67.tar.bz2 \
     |tar xj GenomeAnalysisTK.jar
gatk-register GenomeAnalysisTK.jar # Not gatk3-register here.
rm GenomeAnalysisTK.jar

Install MosaicForecast.

cd /path/to/conda/environment # Optional, any directory would be ok if you set it properly in config.ini
git clone https://github.com/parklab/MosaicForecast.git

Then, you should checkout the specific revision (63d8e60) as following:

cd MosaicForecast
git checkout 63d8e60

Downloading resources

Download all required resource files including the human reference sequences. This step would take some time to complete.

cd /path/to/pipeline
./download_resources.sh

The following reference and resource data will be downloaded from the GATK resource bundle repository.

The following resource files for variant filtering will be downloaded from our github repository.

Usage

You don't need to manually activate the conda environment before running the pipeline. It will be taken care of by the pipeline. All commands below should be running in the directory where you want to get results.

Configuring pipeline

If you changed any locations of tools or resources, you need to set them properly in following config files for each reference genome.

/path/to/pipeline/config.{b37,h19,h38}.ini

sample_list.txt format

The lines starting with # will be commented out and ignored. If you have fastq files,

#sample_id    file_name                       location (full path)
FVLT          FVLT_S15_L003_R1_001.fastq.gz   /path/to/FVLT_S15_L003_R1_001.fastq.gz
FVLT          FVLT_S15_L003_R2_001.fastq.gz   /path/to/FVLT_S15_L003_R2_001.fastq.gz

If you have cram (or bam) files,

#sample_id    file_name       location (full path)
AN02255       AN02255.cram    /path/to/AN02255.cram

Genome mapping

Align fastq files to a reference genome to make a aligned cram, an ummapped bam and flagstats.

python3 /path/to/pipeline/jobs/run_genome_mapping.py \
        -q your_queue \
        --sample-list /path/to/sample_list.txt

If you are going to use the frozen conda environment, you need to set -n bp_frozen.

options

-q, --queue        specify the SGE queue for jobs to be submitted.
-n, --conda-env    specify the name of conda environment (default: bp)
-t, --target-seq   enable targeted sequencing mode to skip mark duplication.
-f, --align-fmt    specify alignment format (cram|bam). Default is cram.
-r, --reference    specify reference genome (b37|hg19|hg38). Default is b37 (GRCh37).
--sample-list      specify sample_list.txt file
-p, --run-gatk-hc  once alignment complete, run the variant calling with the given ploidy options.
--run-filters      once variant calling complete, run the varinat filtering as well.

Variant calling

If you've already done aligning, you can run from the variant calling pipeline. Given the BAM file, run the GATK4 HaplotypeCaller with the given ploidy options.

python3 /path/to/pipeline/jobs/run_variant_calling.py \
        -q your_queue \
        -p 2 12 50 \
        --sample-list /path/to/sample_list.txt

If you are going to use the frozen conda environment, you need to set -n bp_frozen.

options

-q, --queue        specify the SGE queue for jobs to be submitted.
-n, --conda-env    specify the name of conda environment (default: bp)
-f, --align-fmt    specify alignment format (cram|bam). Default is cram.
-r, --reference    specify reference genome (b37|hg19|hg38). Default is b37 (GRCh37).
--sample-list      specify sample_list.txt file
-p, --run-gatk-hc  specify ploidy options used by GATK.
--run-filters      once variant calling complete, run the varinat filtering as well.

Variant filtering

If you've already done aligning and calling variants, you can run from the variant filtering pipeline. In such case, you need to specify a directory where your existing vcf files are using -v (--vcf-directory) option. VCF and index file names should be formed as follows:

<sample name>.ploidy_<ploidy>.vcf.gz
<sample name>.ploidy_<ploidy>.vcf.gz.tbi

Given the BAM and VCF files, run the variant filtering.

python3 /path/to/pipeline/jobs/run_variant_filtering.py \
        -q your_queue \
        -p 50 \
        --sample-list /path/to/sample_list.txt

options

-q, --queue          specify the SGE queue for jobs to be submitted.
-n, --conda-env      specify the name of conda environment (default: bp)
-f, --align-fmt      specify alignment format (cram|bam). Default is cram.
-r, --reference      specify reference genome (b37|hg19|hg38). Default is b37 (GRCh37).
--sample-list        specify sample_list.txt file
-p, --run-gatk-hc    specify ploidy options used by GATK.
--skip-cnvnator      skip the CNV filering
-v, --vcf-directory  If you have VCF files elsewhere, specify the directory where VCF files are.

Contributing

The master branch is protected. To make introduce changes:

  1. Fork this repository
  2. Open a branch with your github username and a short descriptive statement (like kdaily-update-readme). If there is an open issue on this repository, name your branch after the issue (like kdaily-issue-7).
  3. Open a pull request and request a review.