PDXToolkit

PDXToolkit is a mouse reads filtering pipeline from WGS/WES and RNA-seq data

Workflow

Install

• Java (JRE1.8.x)
• conda install -c bioconda bwa=0.7.15
• conda install -c bioconda samtools=1.5
• conda install -c bioconda picard=2.17.11
• conda install -c bioconda bamtools=2.4.0
• conda install -c bioconda ngs-disambiguate=2016.11.10
• conda install -c bioconda star=2.5.4a

Usage

    Version 1.0
    Usage:  pdx_disam_kit.sh -p <command> [options]

Key commands:

    dnaFull            Do full step from *.fq.gz to create new bam 
    rnaFull            Do full step from *.fq.gz to create new *.fq.gz
    
    humanBam           Separate step - Only create human bam
    mouseBam           Separate step - Only create mouse bam
    disambiguate       Separate step - Do disambiguate

Options:

    -p  <command>      Key command
    -n  <name>         Any name for building folder
    -1  <fastq.gz>     Fastq gzip file
    -2  <fastq.gz>     Fastq gzip file
    -o  <directory>    Output directory
    -d  <directory>    Directory. It use to separate step.

Example

1. WGS/WES full step

    sh pdx_disam_kit.sh -p dnaFull -n sample -1 fq1.gz -2 fq2.gz -o outdir
   

2. RNA full step

    sh pdx_disam_kit.sh -p rnaFull -n sample -1 fq1.gz -2 fq2.gz -o outdir
   

3. Separate steps

   • Make human to bam

     sh pdx_disam_kit.sh -p humanBam -n sample -1 fq1.gz -2 fq2.gz -o outdir

   • Make mouse to bam

     sh pdx_disam_kit.sh -p mouseBam -n sample -1 fq1.gz -2 fq2.gz -o outdir

   • Disambiguate

     sh pdx_disam_kit.sh -p disambiguate -n sample -d pdx_wxs_dir

Input

• WGS/WES/RNA-seq

  Pair fastq gzip files (e.g. *.1.fq.gz / *.2.fq.gz)
  

Output

• WGS/WES data output

    *.disam.reAlign.remDup.bam
    *.disam.reAlign.remDup.bam.bai
  

• RNA-seq data output

    *.disam.rna-seq.1.fastq.gz
    *.disam.rna-seq.2.fastq.gz
  

Other versions

Docker version

• Docker files (https://github.com/ding-lab/dockers)

• Usage

   * Only Disambiguate
  
       docker pull hsun9/disambiguate
       docker run hsun9/disambiguate ngs_disambiguate --help
  
  
   * Full pipeline of mouse filter (wxs/wgs data) docker image
  
       docker pull hsun9/disambiguateplus
       docker run hsun9/disambiguateplus ngs_disambiguate --help
  

CWL version

The CWL version developed by Matthew Wyczalkowski (https://github.com/ding-lab/MouseTrap2)

MGI version for VCF/BED file based mouse filtering (gmt somatic filter-mouse-bases):

It's not including in PDXToolkit

NOTE:Please refer to "example.mgi.gmt.sh"
     If test the example.mgi.gmt.sh, you must modify the vars including name, outDir and bed.
     The shell script only run in MGI servers.
     
     The *.permissive.out is the final filtered result.

• Test sample

name="example"                   #-- any name
outDir=/example/outFolder        #-- any dir
bed=/example/target.bed          #-- any bed file
hg19=/path/GRCh37-lite.fa        #-- GRCh37
mm10=/path/Mus_musculus.GRCm38.dna_sm.primary_assembly.fa
chain10=/path/hg19ToMm10.over.chain

gmt somatic filter-mouse-bases --chain-file=$chain10 \
  --human-reference=$hg19 \
  --mouse-reference=$mm10 --variant-file=$bed \
  --filtered-file=$outDir/log/$name.mouse.hg19toMm10.out \
  --output-file=$outDir/$name.hg19toMm10.permissive.out --permissive

## target.bed (create based on VCF)
1 121009 121009 C T

## chain download 
http://hgdownload.cse.ucsc.edu/goldenpath/hg19/liftOver/hg19ToMm10.over.chain.gz

Contact

Hua Sun, hua.sun@wustl.edu