Merge branch 'Delattre/JU28_59vs17_SNP-master' into dev to get BWA GATK...

Merge branch 'Delattre/JU28_59vs17_SNP-master' into dev to get BWA GATK sambamba samblaster and bcftools

.gitignore

@@ -2,3 +2,4 @@ nextflow
 .nextflow.log*
 .nextflow/
 work/
+results

results/.gitignore

-*

src/docker_modules/BWA/0.7.17/Dockerfile

+FROM sambamba:0.6.7
+MAINTAINER Laurent Modolo
+
+ENV BWA_VERSION=0.7.17
+ENV SAMBLASTER_VERSION=0.1.24
+
+ENV PACKAGES curl=7.58.0* \
+    ca-certificates=20180409 \
+    build-essential=12.4* \
+    zlib1g-dev=1:1.2.11*
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends ${PACKAGES} && \
+    apt-get clean
+
+RUN curl -k -L https://github.com/lh3/bwa/releases/download/v${BWA_VERSION}/bwa-${BWA_VERSION}.tar.bz2 -o bwa-v${BWA_VERSION}.tar.bz2 && \
+tar xjf bwa-v${BWA_VERSION}.tar.bz2 && \
+cd bwa-${BWA_VERSION}/ && \
+make && \
+cp bwa /usr/bin && \
+cd .. && \
+rm -R bwa-${BWA_VERSION}/
+
+RUN curl -k -L https://github.com/GregoryFaust/samblaster/releases/download/v.${SAMBLASTER_VERSION}/samblaster-v.${SAMBLASTER_VERSION}.tar.gz -o samblaster-v.${SAMBLASTER_VERSION}.tar.gz && \
+tar xvf samblaster-v.${SAMBLASTER_VERSION}.tar.gz && \
+cd samblaster-v.${SAMBLASTER_VERSION}/ && \
+make && \
+cp samblaster /usr/bin && \
+cd .. && \
+rm -R samblaster-v.${SAMBLASTER_VERSION}/

src/docker_modules/BWA/0.7.17/docker_init.sh

+#!/bin/sh
+docker build src/docker_modules/BWA/0.7.17 -t 'bwa:0.7.17'

src/docker_modules/GATK/4.0.8.1/Dockerfile

+FROM broadinstitute/gatk:4.0.8.1
+MAINTAINER Laurent Modolo
+
+ENV GATK_VERSION=4.0.8.1
+
+RUN cp gatk /usr/bin/

src/docker_modules/GATK/4.0.8.1/docker_init.sh

+#!/bin/sh
+docker build src/docker_modules/GATK/4.0.8.1 -t 'gatk:4.0.8.1'

src/docker_modules/bcftools/1.7/Dockerfile

+FROM ubuntu:18.04
+MAINTAINER Laurent Modolo
+
+ENV BCFTOOLS_VERSION=1.7
+ENV PACKAGES bcftools=${BCFTOOLS_VERSION}*
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends ${PACKAGES} && \
+    apt-get clean

src/docker_modules/bcftools/1.7/docker_init.sh

+#!/bin/sh
+docker build src/docker_modules/bcftools/1.7 -t 'bcftools:1.7'

src/docker_modules/bioawk/1.0/Dockerfile

+FROM ubuntu:18.04
+MAINTAINER Laurent Modolo
+
+ENV BIOAWK_VERSION=1.0
+ENV PACKAGES git=1:2.17* \
+   build-essential=12.4* \
+   ca-certificates=20180409 \
+   zlib1g-dev=1:1.2.11* \
+   byacc
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends ${PACKAGES} && \
+    apt-get clean
+
+RUN git clone https://github.com/lh3/bioawk.git && \
+  cd bioawk && \
+  git checkout tags/v${BIOAWK_VERSION} && \
+  make && \
+  cd .. && \
+  mv bioawk/bioawk /usr/bin/ && \
+  rm -Rf bioawk

src/docker_modules/bioawk/1.0/docker_init.sh

+#!/bin/sh
+docker build src/docker_modules/bioawk/1.0 -t 'bioawk:1.0'

src/docker_modules/sambamba/0.6.7/Dockerfile

+FROM ubuntu:18.04
+MAINTAINER Laurent Modolo
+
+ENV SAMBAMBA_VERSION=0.6.7
+ENV PACKAGES curl=7.58.0* \
+    ca-certificates=20180409 \
+    build-essential=12.4* \
+    zlib1g-dev=1:1.2.11*
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends ${PACKAGES} && \
+    apt-get clean
+
+RUN curl -k -L https://github.com/biod/sambamba/releases/download/v${SAMBAMBA_VERSION}/sambamba_v${SAMBAMBA_VERSION}_linux.tar.bz2 -o sambamba_v${SAMBAMBA_VERSION}_linux.tar.bz2 && \
+tar xvjf sambamba_v${SAMBAMBA_VERSION}_linux.tar.bz2 && \
+mv sambamba /usr/bin/ && \
+rm -R sambamba_v${SAMBAMBA_VERSION}_linux*

src/docker_modules/sambamba/0.6.7/docker_init.sh

+#!/bin/sh
+docker build src/docker_modules/sambamba/0.6.7 -t 'sambamba:0.6.7'

src/docker_modules/samblaster/0.1.24/Dockerfile

+FROM ubuntu:18.04
+MAINTAINER Laurent Modolo
+
+ENV SAMBLASTER_VERSION=0.1.24
+ENV PACKAGES curl=7.58.0* \
+    ca-certificates=20180409 \
+    build-essential=12.4* \
+    zlib1g-dev=1:1.2.11*
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends ${PACKAGES} && \
+    apt-get clean
+
+RUN curl -k -L https://github.com/GregoryFaust/samblaster/releases/download/v.${SAMBLASTER_VERSION}/samblaster-v.${SAMBLASTER_VERSION}.tar.gz -o samblaster-v.${SAMBLASTER_VERSION}.tar.gz && \
+tar xvf samblaster-v.${SAMBLASTER_VERSION}.tar.gz && \
+cd samblaster-v.${SAMBLASTER_VERSION}/ && \
+make && \
+cp samblaster /usr/bin && \
+cd .. && \
+rm -R samblaster-v.${SAMBLASTER_VERSION}/

src/docker_modules/samblaster/0.1.24/docker_init.sh

+#!/bin/sh
+docker build src/docker_modules/samblaster/0.1.24 -t 'samblaster:0.1.24'

src/nf_modules/BWA/indexing.config

+profiles {
+  docker {
+    docker.temp = 'auto'
+    docker.enabled = true
+    process {
+      $index_fasta {
+        container = "bwa:0.7.17"
+      }
+    }
+  }
+  sge {
+    process{
+      $index_fasta {
+        beforeScript = "module purge; module load BWA/0.7.17"
+        executor = "sge"
+        cpus = 1
+        memory = "5GB"
+        time = "6h"
+        queueSize = 1000
+        pollInterval = '60sec'
+        queue = 'h6-E5-2667v4deb128'
+        penv = 'openmp8'
+      }
+    }
+  }
+}

src/nf_modules/BWA/indexing.nf

+params.fasta = "$baseDir/data/bam/*.fasta"
+
+log.info "fasta files : ${params.fasta}"
+
+Channel
+  .fromPath( params.fasta )
+  .ifEmpty { error "Cannot find any bam files matching: ${params.fasta}" }
+  .map { it -> [(it.baseName =~ /([^\.]*)/)[0][1], it]}
+  .set { fasta_file }
+
+process index_fasta {
+  tag "$fasta_id"
+  cpus 4
+  publishDir "results/mapping/index/", mode: 'copy'
+
+  input:
+    set fasta_id, file(fasta) from fasta_file
+
+  output:
+    set fasta_id, "${fasta.baseName}.*" into index_files
+    file "*_bwa_report.txt" into index_files_report
+
+  script:
+"""
+bwa index -p ${fasta.baseName} ${fasta} \
+&> ${fasta.baseName}_bwa_report.txt
+"""
+}
+

src/nf_modules/BWA/mapping_paired.config

+profiles {
+  docker {
+    docker.temp = 'auto'
+    docker.enabled = true
+    process {
+      $mapping_fastq {
+        container = "bwa:0.7.17"
+      }
+    }
+  }
+  sge {
+    process{
+      $mapping_fastq {
+        beforeScript = "module purge; module load BWA/0.7.17"
+        executor = "sge"
+        cpus = 4
+        memory = "5GB"
+        time = "6h"
+        queueSize = 1000
+        pollInterval = '60sec'
+        queue = 'h6-E5-2667v4deb128'
+        penv = 'openmp8'
+      }
+    }
+  }
+}

src/nf_modules/BWA/mapping_paired.nf

+params.fastq = "$baseDir/data/fastq/*_{1,2}.fastq"
+params.index = "$baseDir/data/index/*.index.*"
+
+log.info "fastq files : ${params.fastq}"
+log.info "index files : ${params.index}"
+
+Channel
+  .fromFilePairs( params.fastq )
+  .ifEmpty { error "Cannot find any fastq files matching: ${params.fastq}" }
+  .set { fastq_files }
+Channel
+  .fromPath( params.index )
+  .ifEmpty { error "Cannot find any index files matching: ${params.index}" }
+  .map { it -> [(it.baseName =~ /([^\.]*)/)[0][1], it]}
+  .groupTuple()
+  .set { index_files }
+
+process mapping_fastq {
+  tag "$reads"
+  cpus 4
+  publishDir "results/mapping/sam/", mode: 'copy'
+
+  input:
+  set pair_id, file(reads) from fastq_files
+  set index_id, file(index) from index_files.collect()
+
+  output:
+  file "${pair_id}.sam" into sam_files
+  file "${pair_id}_bwa_report.txt" into mapping_repport_files
+
+  script:
+"""
+bwa mem -t ${task.cpus} \
+${index_id} ${reads[0]} ${reads[1]} \
+-o ${pair_id}.sam &> ${pair_id}_bwa_report.txt
+"""
+}
+

src/nf_modules/BWA/tests.sh

+./nextflow src/nf_modules/BWA/indexing.nf \
+  -c src/nf_modules/BWA/indexing.config \
+  -profile docker \
+  --fasta "data/tiny_dataset/fasta/tiny_v2.fasta"
+
+# ./nextflow src/nf_modules/BWA/mapping_single.nf \
+#   -c src/nf_modules/BWA/mapping_single.config \
+#   -profile docker \
+#   --index "results/mapping/index/tiny_v2.index" \
+#   --fastq "data/tiny_dataset/fastq/tiny*_S.fastq"
+
+./nextflow src/nf_modules/BWA/mapping_paired.nf \
+  -c src/nf_modules/BWA/mapping_paired.config \
+  -profile docker \
+  --index "results/mapping/index/tiny_v2*" \
+  --fastq "data/tiny_dataset/fastq/tiny*_R{1,2}.fastq"
+

src/nf_modules/sambamba/index_bams.config

+profiles {
+  docker {
+    docker.temp = 'auto'
+    docker.enabled = true
+    process {
+      $index_bam {
+        container = "sambamba:0.6.7"
+      }
+    }
+  }
+  sge {
+    process{
+      $index_bam {
+        beforeScript = "module purge; module load sambamba/0.6.7"
+      }
+    }
+  }
+}