diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4678476f37a0b90bf6573d240ff645145a02c2c0..f50bf65084a3e88c07a8ac6c35c1c7a7b96ed498 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -11,6 +11,13 @@ that users understand how the changes affect the new version.
version 4.0.0-develop
---------------------------
++ Default docker images for bwa, bwakit and hisat2 updated to include samtools
+ 1.10.
++ Alignment tasks (STAR, Hisat2, BWA) now produce BAM files at level 1
+ compression.
++ Hisat2 task has added controls for samtools.
++ Alignment tasks no longer produce BAM indexes as these are not needed
+ by the markduplicates step.
+ Picard Markduplicates now uses 7G of RAM just like in GATK's best practice
example pipeline.
+ Picard SortSam added as a task.
@@ -18,7 +25,7 @@ version 4.0.0-develop
BAM files.
+ Changed PicardMarkduplicates to use COMPRESSION_LEVEL=1 by default with
the htsjdk deflater.
- This makes the task finish in 30% less time at the cost of a 6% larger BAM
+ This makes the task finish in 32% less time at the cost of a 8% larger BAM
file.
+ Added sambamba markdup and sambamba sort. NOTE: samtools sort is more
efficient and is recommended.
diff --git a/bcftools.wdl b/bcftools.wdl
index e1ec3059d022a76be71550ff9e993d8f1bc6d445..b99a8cf581517272beeb174dace0c2ed7587f4ce 100644
--- a/bcftools.wdl
+++ b/bcftools.wdl
@@ -22,23 +22,32 @@ version 1.0
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.
-task Bcf2Vcf {
+task View {
input {
- File bcf
- String outputPath = "./bcftools/SV.vcf"
- String memory = "2G"
- Int timeMinutes = 1 + ceil(size(bcf, "G"))
- String dockerImage = "quay.io/biocontainers/bcftools:1.9--ha228f0b_3"
+ File inputFile
+ String outputPath = "output.vcf.gz"
+ String memory = "256M"
+ Int timeMinutes = 1 + ceil(size(inputFile, "G"))
+ String dockerImage = "quay.io/biocontainers/bcftools:1.10.2--h4f4756c_2"
+ String outputType = "z"
+ Int compressionLevel = 1
}
command {
set -e
mkdir -p "$(dirname ~{outputPath})"
- bcftools view ~{bcf} -O v -o ~{outputPath}
+ bcftools view \
+ -o ~{outputPath} \
+ -O ~{outputType} \
+ -l ~{compressionLevel} \
+ ~{inputFile}
+ bcftools index --tbi ~{outputPath}
+
}
output {
File outputVcf = outputPath
+ File outputVcfIndex = outputPath + ".tbi"
}
runtime {
@@ -48,9 +57,11 @@ task Bcf2Vcf {
}
parameter_meta {
- bcf: {description: "The generated BCF from an SV caller", category: "required"}
+ inputFile: {description: "A vcf or bcf file", category: "required"}
outputPath: {description: "The location the output VCF file should be written.", category: "common"}
+ outputType: {description: "Output type: v=vcf, z=vcf.gz, b=bcf, u=uncompressed bcf"}
memory: {description: "The amount of memory this job will use.", category: "advanced"}
+ compressionLevel: {description: "Compression level from 0 (uncompressed) to 9 (best).", category: "advanced"}
timeMinutes: {description: "The maximum amount of time the job will run in minutes.", category: "advanced"}
dockerImage: {description: "The docker image used for this task. Changing this may result in errors which the developers may choose not to address.", category: "advanced"}
}
@@ -87,7 +98,7 @@ task Stats {
Int threads = 0
Int timeMinutes = 1 + 2* ceil(size(select_all([inputVcf, compareVcf]), "G")) # TODO: Estimate, 2 minutes per GB, refine later.
String memory = "256M"
- String dockerImage = "quay.io/biocontainers/bcftools:1.9--ha228f0b_3"
+ String dockerImage = "quay.io/biocontainers/bcftools:1.10.2--h4f4756c_2"
}
command {
diff --git a/chunked-scatter.wdl b/chunked-scatter.wdl
index 3ef0c74784c8fb91e78ecb2055dfa84be9174c88..b54a7d2e909a20d43bbe194d9f28496b7753ce5d 100644
--- a/chunked-scatter.wdl
+++ b/chunked-scatter.wdl
@@ -30,7 +30,7 @@ task ChunkedScatter {
String memory = "256M"
Int timeMinutes = 2
- String dockerImage = "quay.io/biocontainers/chunked-scatter:0.2.0--py_0"
+ String dockerImage = "quay.io/biocontainers/chunked-scatter:1.0.0--py_0"
}
command {
diff --git a/cutadapt.wdl b/cutadapt.wdl
index ad32ff21aff4958f9955628753a7b68468a855cd..d125af43b3039fd2ba9aa2b63b0819d4676fbf41 100644
--- a/cutadapt.wdl
+++ b/cutadapt.wdl
@@ -81,7 +81,7 @@ task Cutadapt {
Int cores = 4
String memory = "~{300 + 100 * cores}M"
Int timeMinutes = 1 + ceil(size([read1, read2], "G") * 12.0 / cores)
- String dockerImage = "quay.io/biocontainers/cutadapt:2.8--py37h516909a_0"
+ String dockerImage = "quay.io/biocontainers/cutadapt:2.10--py37h516909a_0"
}
String realRead2output = select_first([read2output, "cut_r2.fq.gz"])
diff --git a/gatk.wdl b/gatk.wdl
index f20bf70ad54a61de7123fe48e80bdbf538e3e742..e0209a0c4ef0f4b19fde26f3b47f0d0bd2b33fe0 100644
--- a/gatk.wdl
+++ b/gatk.wdl
@@ -35,7 +35,7 @@ task AnnotateIntervals {
String memory = "3G"
String javaXmx = "2G"
Int timeMinutes = 5
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -96,7 +96,7 @@ task ApplyBQSR {
Int memoryMb = javaXmxMb + 512
Int javaXmxMb = 2048
Int timeMinutes = 120 # This will likely be used with intervals, as such size based estimation can't be used.
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -168,7 +168,7 @@ task BaseRecalibrator {
Int memoryMb = javaXmxMb + 512
Int javaXmxMb = 1024
Int timeMinutes = 120 # This will likely be used with intervals, as such size based estimation can't be used.
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -227,7 +227,7 @@ task CalculateContamination {
String memory = "13G"
String javaXmx = "12G"
Int timeMinutes = 180
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.2.0--1"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -271,7 +271,7 @@ task CallCopyRatioSegments {
String memory = "3G"
String javaXmx = "2G"
Int timeMinutes = 2
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -320,7 +320,7 @@ task CollectAllelicCounts {
String memory = "11G"
String javaXmx = "10G"
Int timeMinutes = 120
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -376,7 +376,7 @@ task CollectReadCounts {
String memory = "8G"
String javaXmx = "7G"
Int timeMinutes = 1 + ceil(size(inputBam, "G") * 5)
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -433,7 +433,7 @@ task CombineGVCFs {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 1 + ceil(size(gvcfFiles, "G") * 8)
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -607,7 +607,7 @@ task DenoiseReadCounts {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 5
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -665,7 +665,7 @@ task FilterMutectCalls {
String memory = "13G"
String javaXmx = "12G"
Int timeMinutes = 60
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.2.0--1"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -728,7 +728,7 @@ task GatherBqsrReports {
Int memoryMb = 256 + javaXmxMb
Int javaXmxMb = 256
Int timeMinutes = 1
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -774,7 +774,7 @@ task GenomicsDBImport {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 180
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -833,7 +833,7 @@ task GenotypeGVCFs {
String memory = "7G"
String javaXmx = "6G"
Int timeMinutes = 120 # This will likely be used with intervals, as such size based estimation can't be used.
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -899,7 +899,7 @@ task GetPileupSummaries {
String memory = "13G"
String javaXmx = "12G"
Int timeMinutes = 120
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.2.0--1"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -966,7 +966,7 @@ task HaplotypeCaller {
# Memory increases with time used. 4G should cover most use cases.
Int javaXmxMb = 4096
Int timeMinutes = 400 # This will likely be used with intervals, as such size based estimation can't be used.
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1040,7 +1040,7 @@ task LearnReadOrientationModel {
String memory = "13G"
String javaXmx = "12G"
Int timeMinutes = 120
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.2.0--1"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1079,7 +1079,7 @@ task MergeStats {
String memory = "15G"
String javaXmx = "14G"
Int timeMinutes = 30
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1126,7 +1126,7 @@ task ModelSegments {
String memory = "11G"
String javaXmx = "10G"
Int timeMinutes = 60
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1202,7 +1202,7 @@ task MuTect2 {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 240
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1389,7 +1389,7 @@ task PreprocessIntervals {
String memory = "4G"
String javaXmx = "3G"
Int timeMinutes = 1 + ceil(size(referenceFasta, "G") * 6)
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1447,7 +1447,7 @@ task SelectVariants {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 60
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1507,7 +1507,7 @@ task SplitNCigarReads {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 120 # This will likely be used with intervals, as such size based estimation can't be used.
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1576,7 +1576,7 @@ task VariantEval {
String javaXmx = "4G"
# TODO: Refine estimate. For now 4 minutes per GB of input.
Int timeMinutes = ceil(size(flatten([evalVcfs, comparisonVcfs, select_all([referenceFasta, dbsnpVCF])]), "G") * 20)
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.7.0--py38_0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
@@ -1647,7 +1647,7 @@ task VariantFiltration {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 120
- String dockerImage = "quay.io/biocontainers/gatk4:4.1.0.0--0"
+ String dockerImage = "quay.io/biocontainers/gatk4:4.1.8.0--py38h37ae868_0"
}
command {
diff --git a/hisat2.wdl b/hisat2.wdl
index 5937f86d433fd4be82771bc6214e0779fbe48e95..c24610edc6fdc732a70040756b62483d7dc509d4 100644
--- a/hisat2.wdl
+++ b/hisat2.wdl
@@ -34,12 +34,15 @@ task Hisat2 {
String summaryFilePath = basename(outputBam, ".bam") + ".summary.txt"
Int threads = 4
- String memory = "~{threads + 5 + ceil(size(indexFiles, "G"))}G"
+ Int sortThreads = 1
+ Int sortMemoryPerThreadGb = 2
+ Int compressionLevel = 1
+ Int memoryGb = 1 + threads + ceil(size(indexFiles, "G") * 1.2) + sortMemoryPerThreadGb * sortThreads
Int timeMinutes = 1 + ceil(size([inputR1, inputR2], "G") * 180 / threads)
# quay.io/biocontainers/mulled-v2-a97e90b3b802d1da3d6958e0867610c718cb5eb1
# is a combination of hisat2 and samtools
- # hisat2=2.1.0, samtools=1.8
- String dockerImage = "quay.io/biocontainers/mulled-v2-a97e90b3b802d1da3d6958e0867610c718cb5eb1:2388ff67fc407dad75774291ca5038f40cac4be0-0"
+ # hisat2=2.2.0, samtools=1.10
+ String dockerImage = "quay.io/biocontainers/mulled-v2-a97e90b3b802d1da3d6958e0867610c718cb5eb1:2880dd9d8ad0a7b221d4eacda9a818e92983128d-0"
}
String bamIndexPath = sub(outputBam, "\.bam$", ".bai")
@@ -59,18 +62,21 @@ task Hisat2 {
~{true="--dta" false="" downstreamTranscriptomeAssembly} \
--new-summary \
--summary-file ~{summaryFilePath} \
- | samtools sort > ~{outputBam}
- samtools index ~{outputBam} ~{bamIndexPath}
+ | samtools sort \
+ ~{"-@ " + sortThreads} \
+ -m ~{sortMemoryPerThreadGb}G \
+ -l ~{compressionLevel} \
+ - \
+ -o ~{outputBam}
}
output {
File bamFile = outputBam
- File bamIndex = bamIndexPath
File summaryFile = summaryFilePath
}
runtime {
- memory: memory
+ memory: "~{memoryGb}G"
cpu: threads + 1
time_minutes: timeMinutes
docker: dockerImage
@@ -88,9 +94,12 @@ task Hisat2 {
downstreamTranscriptomeAssembly: {description: "Equivalent to hisat2's `--dta` flag.", category: "advanced"}
summaryFilePath: {description: "Where the summary file should be written.", category: "advanced"}
threads: {description: "The number of threads to use.", category: "advanced"}
- memory: {description: "The amount of memory this job will use.", category: "advanced"}
+ memoryGb: {description: "The amount of memory this job will use in gigabytes.", category: "advanced"}
timeMinutes: {description: "The maximum amount of time the job will run in minutes.", category: "advanced"}
dockerImage: {description: "The docker image used for this task. Changing this may result in errors which the developers may choose not to address.",
category: "advanced"}
+ sortThreads: {description: "The number of threads to use for sorting.", category: "advanced"}
+ sortMemoryPerThreadGb: {description: "The amount of memory for each sorting thread in gigabytes.", category: "advanced"}
+ compressionLevel: {description: "The compression level of the output BAM.", category: "advanced"}
}
}
\ No newline at end of file
diff --git a/htseq.wdl b/htseq.wdl
index 35faeef3a1158634cfdbff45f663bab9d40d035e..cbd8e2acded8d7e89157201a548aa6a29768096b 100644
--- a/htseq.wdl
+++ b/htseq.wdl
@@ -25,23 +25,23 @@ task HTSeqCount {
Array[File]+ inputBams
File gtfFile
String outputTable = "output.tsv"
- String format = "bam"
String order = "pos"
String stranded = "no"
String? featureType
String? idattr
Array[String] additionalAttributes = []
- String memory = "40G"
+ Int nprocesses = 1
+ String memory = "8G"
Int timeMinutes = 10 + ceil(size(inputBams, "G") * 60)
- String dockerImage = "quay.io/biocontainers/htseq:0.11.2--py37h637b7d7_1"
+ String dockerImage = "quay.io/biocontainers/htseq:0.12.4--py37hb3f55d8_0"
}
command {
set -e
mkdir -p "$(dirname ~{outputTable})"
htseq-count \
- -f ~{format} \
+ --nprocesses ~{nprocesses} \
-r ~{order} \
-s ~{stranded} \
~{"--type " + featureType} \
@@ -49,7 +49,7 @@ task HTSeqCount {
~{true="--additional-attr " false="" length(additionalAttributes) > 0 }~{sep=" --additional-attr " additionalAttributes} \
~{sep=" " inputBams} \
~{gtfFile} \
- > ~{outputTable}
+ -c ~{outputTable}
}
output {
@@ -57,6 +57,7 @@ task HTSeqCount {
}
runtime {
+ cpu: nprocesses
time_minutes: timeMinutes
memory: memory
docker: dockerImage
@@ -66,7 +67,7 @@ task HTSeqCount {
inputBams: {description: "The input BAM files.", category: "required"}
gtfFile: {description: "A GTF/GFF file containing the features of interest.", category: "required"}
outputTable: {description: "The path to which the output table should be written.", category: "common"}
- format: {description: "Equivalent to the -f option of htseq-count.", category: "advanced"}
+ nprocesses: {description: "Number of processes to run htseq with.", category: "advanced"}
order: {description: "Equivalent to the -r option of htseq-count.", category: "advanced"}
stranded: {description: "Equivalent to the -s option of htseq-count.", category: "common"}
featureType: {description: "Equivalent to the --type option of htseq-count.", category: "advanced"}
diff --git a/picard.wdl b/picard.wdl
index 4dd4d970d7c48cb99ed4b948c6c4ef8313bcc555..adb55b4b52d2dce2bc3ae45a67a02b479dd9c181 100644
--- a/picard.wdl
+++ b/picard.wdl
@@ -89,7 +89,7 @@ task CollectMultipleMetrics {
Int javaXmxMb = 3072
# Additional * 2 because picard multiple metrics reads the reference fasta twice.
Int timeMinutes = 1 + ceil(size(referenceFasta, "G") * 3 * 2) + ceil(size(inputBam, "G") * 6)
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
@@ -205,7 +205,7 @@ task CollectRnaSeqMetrics {
String javaXmx = "8G"
# With 6 minutes per G there were several timeouts.
Int timeMinutes = 1 + ceil(size(inputBam, "G") * 12)
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
command {
@@ -263,7 +263,7 @@ task CollectTargetedPcrMetrics {
String memory = "4G"
String javaXmx = "3G"
Int timeMinutes = 1 + ceil(size(inputBam, "G") * 6)
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
command {
@@ -322,7 +322,7 @@ task CreateSequenceDictionary {
String memory = "3G"
String javaXmx = "2G"
- String dockerImage = "quay.io/biocontainers/picard:2.22.3--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
command {
@@ -370,7 +370,7 @@ task GatherBamFiles {
Boolean createMd5File = false
# One minute per input gigabyte.
Int timeMinutes = 1 + ceil(size(inputBams, "G") * 1)
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
command {
@@ -422,7 +422,7 @@ task GatherVcfs {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 1 + ceil(size(inputVcfs, "G") * 2)
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
command {
@@ -469,6 +469,7 @@ task MarkDuplicates {
Boolean createMd5File = false
Boolean useJdkInflater = true # Slightly faster than the intel one.
# Better results for compression level 1 (much smaller). Higher compression levels similar to intel deflater.
+ # NOTE: this might change in the future when the intel deflater is updated!
Boolean useJdkDeflater = true
# In GATK Best practices pipeline MarkDuplicates is given a 7G VM.
@@ -477,7 +478,7 @@ task MarkDuplicates {
String memoryMb = javaXmxMb + 512
Int timeMinutes = 1 + ceil(size(inputBams, "G") * 8)
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
# The program default for READ_NAME_REGEX is appropriate in nearly every case.
# Sometimes we wish to supply "null" in order to turn off optical duplicate detection
@@ -530,7 +531,10 @@ task MarkDuplicates {
outputBamPath: {description: "The location where the ouptut BAM file should be written.", category: "required"}
metricsPath: {description: "The location where the output metrics file should be written.", category: "required"}
read_name_regex: {description: "Equivalent to the `READ_NAME_REGEX` option of MarkDuplicates.", category: "advanced"}
-
+ createMd5File: {description: "Whether to create a md5 file for the created BAM file.", category: "advanced"}
+ useJdkInflater: {description: "True, uses the java inflater. False, uses the optimized intel inflater.", category: "advanced"}
+ useJdkDeflater: {description: "True, uses the java deflator to compress the BAM files. False uses the optimized intel deflater.", category: "advanced"}
+ compressionLevel: {description: "The compression level at which the BAM files are written", category: "advanced"}
memoryMb: {description: "The amount of memory this job will use in megabytes.", category: "advanced"}
javaXmxMb: {description: "The maximum memory available to the program in megabytes. Should be lower than `memoryMb` to accommodate JVM overhead.",
category: "advanced"}
@@ -550,7 +554,7 @@ task MergeVCFs {
String memory = "5G"
String javaXmx = "4G"
Int timeMinutes = 1 + ceil(size(inputVCFs, "G"))
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
# Using MergeVcfs instead of GatherVcfs so we can create indices
@@ -599,7 +603,7 @@ task SamToFastq {
String memory = "17G"
String javaXmx = "16G" # High memory default to avoid crashes.
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
File? NONE
}
@@ -636,7 +640,7 @@ task ScatterIntervalList {
String memory = "4G"
String javaXmx = "3G"
- String dockerImage = "quay.io/biocontainers/picard:2.20.5--0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
command {
@@ -677,9 +681,7 @@ task SortSam {
# GATK Best practices uses 75000 here: https://github.com/gatk-workflows/broad-prod-wgs-germline-snps-indels/blob/d2934ed656ade44801f9cfe1c0e78d4f80684b7b/PairedEndSingleSampleWf-fc-hg38.wdl#L778
Int XmxGb = ceil(maxRecordsInRam / 125001.0)
Int timeMinutes = 1 + ceil(size(inputBam, "G") * 3)
- # A mulled container is needed to have both picard and bwa in one container.
- # This container contains: picard (2.18.7), bwa (0.7.17-r1188)
- String dockerImage = "quay.io/biocontainers/picard:2.23.1--h37ae868_0"
+ String dockerImage = "quay.io/biocontainers/picard:2.23.2--0"
}
command {
diff --git a/samtools.wdl b/samtools.wdl
index 5daf57ab3375ddb57193b4ad868268c755ddd079..0b8394bff308d65874d03170918569417b1c751e 100644
--- a/samtools.wdl
+++ b/samtools.wdl
@@ -114,7 +114,7 @@ task Fastq {
Int threads = 1
String memory = "1G"
Int timeMinutes = 1 + ceil(size(inputBam) * 2)
- String dockerImage = "quay.io/biocontainers/samtools:1.8--h46bd0b3_5"
+ String dockerImage = "quay.io/biocontainers/samtools:1.10--h9402c20_2"
}
command {
@@ -170,7 +170,7 @@ task FilterShortReadsBam {
String outputPathBam
String memory = "1G"
Int timeMinutes = 1 + ceil(size(bamFile, "G") * 8)
- String dockerImage = "quay.io/biocontainers/samtools:1.8--h46bd0b3_5"
+ String dockerImage = "quay.io/biocontainers/samtools:1.10--h9402c20_2"
}
String outputPathBamIndex = sub(outputPathBam, "\.bam$", ".bai")
@@ -211,7 +211,7 @@ task Flagstat {
String memory = "256M" # Only 40.5 MiB used for 150G bam file.
Int timeMinutes = 1 + ceil(size(inputBam, "G"))
- String dockerImage = "quay.io/biocontainers/samtools:1.8--h46bd0b3_5"
+ String dockerImage = "quay.io/biocontainers/samtools:1.10--h9402c20_2"
}
command {
@@ -247,7 +247,7 @@ task Index {
String? outputBamPath
String memory = "2G"
Int timeMinutes = 1 + ceil(size(bamFile, "G") * 4)
- String dockerImage = "quay.io/biocontainers/samtools:1.8--h46bd0b3_5"
+ String dockerImage = "quay.io/biocontainers/samtools:1.10--h9402c20_2"
}
# Select_first is needed, otherwise womtool validate fails.
@@ -296,7 +296,7 @@ task Markdup {
String outputBamPath
Int timeMinutes = 1 + ceil(size(inputBam, "G") * 2)
- String dockerImage = "quay.io/biocontainers/samtools:1.8--h46bd0b3_5"
+ String dockerImage = "quay.io/biocontainers/samtools:1.10--h9402c20_2"
}
command {
@@ -332,7 +332,7 @@ task Merge {
Int threads = 1
Int timeMinutes = 1 + ceil(size(bamFiles, "G") * 2)
- String dockerImage = "quay.io/biocontainers/samtools:1.8--h46bd0b3_5"
+ String dockerImage = "quay.io/biocontainers/samtools:1.10--h9402c20_2"
}
String indexPath = sub(outputBamPath, "\.bam$",".bai")
@@ -483,7 +483,7 @@ task View {
Int threads = 1
String memory = "1G"
Int timeMinutes = 1 + ceil(size(inFile, "G") * 5)
- String dockerImage = "quay.io/biocontainers/samtools:1.8--h46bd0b3_5"
+ String dockerImage = "quay.io/biocontainers/samtools:1.10--h9402c20_2"
}
String outputIndexPath = basename(outputFileName) + ".bai"
diff --git a/star.wdl b/star.wdl
index 4da67f72609c3bd27d03301c3cd0ed39c9da4a99..3d0e2eb0e7c0fcfd0518103c962eed2436ee9422 100644
--- a/star.wdl
+++ b/star.wdl
@@ -103,6 +103,7 @@ task Star {
String? twopassMode = "Basic"
Array[String]? outSAMattrRGline
String? outSAMunmapped = "Within KeepPairs"
+ Int outBAMcompression = 1
Int? limitBAMsortRAM
Int runThreadN = 4
@@ -129,6 +130,7 @@ task Star {
--outFileNamePrefix ~{outFileNamePrefix} \
--genomeDir ~{sub(indexFiles[0], basename(indexFiles[0]), "")} \
--outSAMtype ~{outSAMtype} \
+ --outBAMcompression ~{outBAMcompression} \
--readFilesCommand ~{readFilesCommand} \
~{"--outFilterScoreMin " + outFilterScoreMin} \
~{"--outFilterScoreMinOverLread " + outFilterScoreMinOverLread} \
@@ -172,6 +174,7 @@ task Star {
limitBAMsortRAM: {description: "Equivalent to star's `--limitBAMsortRAM` option.", category: "advanced"}
runThreadN: {description: "The number of threads to use.", category: "advanced"}
memory: {description: "The amount of memory this job will use.", category: "advanced"}
+ outBAMcompression: {description: "The compression level of the output BAM.", category: "advanced"}
timeMinutes: {description: "The maximum amount of time the job will run in minutes.", category: "advanced"}
dockerImage: {description: "The docker image used for this task. Changing this may result in errors which the developers may choose not to address.", category: "advanced"}
}