IMPORTANT: This is the legacy GATK documentation. This information is only valid until Dec 31st 2019. For latest documentation and forum click here

created by LeeTL1220

on 2016-01-14

Notice: these workflows are out of date and should no longer be used. See the WDL scripts instead.

These workflows are not officially supported, but are used extensively, internally, for our evaluations.

These instructions assume some familiarity with running Queue.

Download Queue scala scripts

Download the following files to the same directory. This step only needs to be done once. - CaseSampleHBExomePipeline.scala [http://www.broadinstitute.org/~lichtens/gatk4cnv/CaseSampleHBExomePipeline.scala] - CreatePoNPipeline.scala [http://www.broadinstitute.org/~lichtens/gatk4cnv/CreatePoNPipeline.scala] - recapseg-hb-eval.jar (Queue) [http://www.broadinstitute.org/~lichtens/gatk4cnv/recapseg-hb-eval.jar] - gatk-protected.jar [http://www.broadinstitute.org/~lichtens/gatk4cnv/gatk-protected.jar]

Please note: These locations are temporary and will be updated in the future.

Test that invoking the help documentation works

java -jar recapseg-hb-eval.jar -S CreatePoNPipeline.scala --help

java -jar recapseg-hb-eval.jar -S CaseSampleHBExomePipeline.scala --help

Example case sample run with ACNV

Parameters to the recapseg-hb-eval.jar may need to be adjusted depending on your execution environment

```

!/bin/bash -l

. /broad/tools/scripts/useuse use .hdfview-2.9 use Java-1.8 use .r-3.1.3-gatk-only

Modify parameters below this line

Two simple text files listing each bam file on a separate line

Note that the lines in each file must correspond to case-control pairs.

CASESAMPLES=tumorbamfiles.txt CONTROLSAMPLES=normalbamfiles.txt

Output location (absolute paths recommended).

Seg file will appear in $OUTDIR/tumor_pcov

Calls will appear in $OUTDIR/caller

OUTDIR=/home/username/evals/out_case/

Downloaded gatk-protected.jar

GATK4PJAR=gatk-protected.jar

Directory of HDF5 JNI shared libraries (.so/.dynlib)

HDFLOC=/broad/software/free/Linux/redhat6x8664/pkgs/hdfview2.9/HDFView/lib/linux/

Reference (This is b37 at the Broad Institute)

REF=/seq/references/Homosapiensassembly19/v1/Homosapiensassembly19.fasta

Target list. Must be the same as was used to create the PoN.

TARGETS=/home/lichtens/mytargetlist.bed

PoN file must be created with the same target file (above) and settings.

PON=/home/lichtens/evals/outpon/createpon/mybloodnormals.pon

How much memory to allocate to each job, in GB

MEM=8

List of SNPs in the interval_list format

SNPLIST=mylistofcommonhetsites.interval_list

See --help for descriptions of these parameters.

If -keepDups was specified for the PoN, it must be specified for case samples as well.

If -pd 250 was specified for the PoN, it must be specified for case samples as well.

-rawcov will generate a separate file for the raw counts as a separate process.

PD=250 OTHEROPTS=" -jobResReq virtualfree=${MEM}G -keepDups -rawcov -jobQueue gsa -pd ${PD} -noWt " OTHEROPTSACNV=" -acnv -snp ${SNPLIST} -icontrol ${CONTROLSAMPLES} "

Do not modify below this line

Run the sample(s)

java -jar recapseg-hb-eval.jar -S CaseSampleHBExomePipeline.scala -mem ${MEM} -pon ${PON} -i ${CASESAMPLES} -o ${OUTDIR} -gatk4pjar ${HBJAR} -r ${REF} -L ${TARGETS} -qsub -run -logDir ${OUTDIR} -hvl ${HDFLOC} -Z ${Z} ${OTHEROPTS} ${OTHEROPTSACNV}

```

Example case sample run without ACNV

Parameters to the recapseg-hb-eval.jar may need to be adjusted depending on your execution environment

```

!/bin/bash -l

. /broad/tools/scripts/useuse use .hdfview-2.9 use Java-1.8 use .r-3.1.3-gatk-only

Modify parameters below this line

A simple text file listing each bam file on a separate line

INPUTBAMS=casebam_list.txt

Output location (absolute paths recommended).

Seg file will appear in $OUTDIR/tumor_pcov

Calls will appear in $OUTDIR/caller

OUTDIR=/home/username/evals/out_case/

Downloaded gatk-protected.jar

GATK4PJAR=gatk-protected.jar

Directory of HDF5 JNI shared libraries (.so/.dynlib)

HDFLOC=/broad/software/free/Linux/redhat6x8664/pkgs/hdfview2.9/HDFView/lib/linux/

Reference (This is b37 at the Broad Institute)

REF=/seq/references/Homosapiensassembly19/v1/Homosapiensassembly19.fasta

Target list. Must be the same as was used to create the PoN.

TARGETS=/home/lichtens/mytargetlist.bed

PoN file must be created with the same target file (above) and settings.

PON=/home/lichtens/evals/outpon/createpon/mybloodnormals.pon

How much memory to allocate to each job, in GB

MEM=8

See --help for descriptions of these parameters.

If -keepDups was specified for the PoN, it must be specified for case samples as well.

If -pd 250 was specified for the PoN, it must be specified for case samples as well.

-rawcov will generate a separate file for the raw counts as a separate process.

PD=250 OTHEROPTS=" -jobResReq virtualfree=${MEM}G -keepDups -rawcov -jobQueue gsa -pd ${PD} -noWt "

Do not modify below this line

Run the sample(s)

java -jar recapseg-hb-eval.jar -S CaseSampleHBExomePipeline.scala -mem ${MEM} -pon ${PON} -i ${INPUTBAMS} -o ${OUTDIR} -hbJar ${GATK4PJAR} -r ${REF} -L ${TARGETS} -qsub -run -logDir ${OUTDIR} -hvl ${HDFLOC} ${OTHEROPTS}

```

Example create PoN run

Parameters to the Queue jar (recapseg-hb-eval.jar) may need to be adjusted depending on your execution environment

```

!/bin/bash -l

. /broad/tools/scripts/useuse use .hdfview-2.9 use Java-1.8 use .r-3.1.3-gatk-only

Modify parameters below this line

A simple text file listing each bam file on a separate line

INPUTBAMS=bloodnormalsbamlist.txt

Output location (absolute paths recommended).

PoN file will appear as $OUTDIR/createpon/$PONFILENAME

OUTDIR=/home/lichtens/evals/out_pon/

Downloaded gatk-protected.jar

GATK4PJAR=gatk-protected.jar

Directory of HDF5 JNI shared libraries (.so/.dynlib)

HDFLOC=/broad/software/free/Linux/redhat6x8664/pkgs/hdfview2.9/HDFView/lib/linux/

Reference (This is b37 at the Broad Institute)

REF=/seq/references/Homosapiensassembly19/v1/Homosapiensassembly19.fasta

Base PoN filename

This parameter is only the base filename. Do not specify a directory. This file will appear as

$OUTDIR/createpon/$PONFILENAME

PONFILENAME=myblood_normals.pon

Target list that was used in the capture process.

TARGETS=/home/lichtens/mytargetlist.bed

How much memory to allocate to each coverage job, in GB.

208k targets needs approx. 6GB RAM

MEM=6

Larger PoNs can require a lot of RAM (208k targets x 300 samples needs approx. 14GB RAM)

MEM_PON=14

Number of cores to use when multicore functionality is available.

CORES=4

See --help for descriptions of these parameters.

If -keepDups was specified for the PoN, it must be specified for case samples as well.

If -pd 250 was specified for the PoN, it must be specified for case samples as well.

-rawcov will generate a separate file for the raw counts as a separate process.

OTHEROPTS=" -jobResReq 'virtualfree=${MEM}G' -keepDups -rawcov -jobQueue gsa -pd 250 -sparkMaster 'local[${CORES}]' "

Do not modify below this line

Run the sample(s)

java -jar recapseg-hb-eval.jar -S CreatePoNPipeline.scala -pon ${PON} -i ${IN} -o ${OUTDIR} -gatk4pjar ${HBJAR} -r ${REF} -L ${TARGETS} -qsub -run -logDir ${OUTDIR} -hvl ${HDFLOC} -mem ${MEM} -mempon ${MEMPON} ${OTHER_OPTS}

```

Updated on 2017-12-29

From LeeTL1220 on 2016-01-20

To run the jobs on a local machine, remove “-qsub” and “-jobQueue gsa”. Add “ -maxConcurrentRun 10 -jobRunner ParallelShell “, where 10 is the number of processes to use.

From Geraldine_VdAuwera on 2017-12-29

Notice: these workflows are out of date and should no longer be used. See the WDL scripts instead.