src/nf_modules/stringtie2/main.nf

+// SPDX-FileCopyrightText: 2022 Laurent Modolo <laurent.modolo@ens-lyon.fr>
+//
+// SPDX-License-Identifier: AGPL-3.0-or-later
+
+version = "2.1.1"
+container_url = "lbmc/stringtie2:${version}"
+
+process jcount {
+  container = "${container_url}"
+  label "big_mem_multi_cpus"
+  tag "$file_id"
+  if (params.salmon_out != "") {
+    publishDir "results/${params.salmon_out}", mode: 'copy'
+  }
+
+  input:
+    tuple val(file_id), path(bam)
+
+  output:
+    tuple val(file_id), path("*.sf"), emit: quant
+  script:
+"""
+salmon quant -l A --noErrorModel -t XXXXXXXXXX -a ${bam} -p 4 -o ${params.salmon_out}
+"""
+}
\ No newline at end of file

src/nf_modules/trinity/main.nf

+// SPDX-FileCopyrightText: 2022 Laurent Modolo <laurent.modolo@ens-lyon.fr>
+//
+// SPDX-License-Identifier: AGPL-3.0-or-later
+
+version = "2.13.2"
+container_url = "lbmc/trinity:${version}"
+
+params.sample = 3
+params.min_glue = 1
+params.min_contig_length = 200
+params.assembly_out = ""
+
+
+workflow assembly {
+  take:
+    fastq
+  main:
+    complete_assembly(fastq)
+    super_transcript(complete_assembly.out.fasta)
+
+  emit:
+    fasta = complete_assembly.out.fasta
+    super_transcript = super_transcript.out.fasta
+}
+
+process complete_assembly {
+  container = "${container_url}"
+  label "big_mem_multi_cpus"
+  tag "$file_id"
+  if (params.assembly_out != "") {
+    publishDir "results/${params.assembly_out}", mode: 'copy'
+  }
+
+  input:
+    tuple val(file_id), path(fastq)
+
+  output:
+    tuple val(file_id), path("trinity_output_${file_prefix}/"), emit: folder
+    tuple val(file_id), path("trinity_output_${file_prefix}.Trinity.fasta"), emit: fasta
+    tuple val(file_id), path("trinity_output_${file_prefix}.Trinity.fasta.gene_trans_map"), emit: gene_map 
+    tuple val(file_id), path("trinity_output_${file_prefix}/salmon_outdir/quant.sf"), emit: quant
+
+  script:
+
+  switch(file_id) {
+    case {it instanceof List}:
+      file_prefix = file_id[0]
+    break;
+    case {it instanceof Map}:
+      file_prefix = file_id.values()[0]
+    break;
+    default:
+      file_prefix = file_id
+    break;
+  };
+  def memory = "${task.memory}" - ~/\s*GB/
+
+  if (fastq.size() == 2)
+"""
+  mkdir trinity_output_${file_prefix}
+  Trinity \
+    --seqType fq \
+    --max_memory ${memory}G \
+    --left ${fastq[0]} \
+    --right ${fastq[1]} \
+    --CPU ${task.cpus} \
+    --min_glue ${params.min_glue} \
+    --min_contig_length ${params.min_contig_length} \
+    --output trinity_output_${file_prefix}
+"""
+  else
+"""
+  mkdir trinity_output_${file_prefix}
+  Trinity \
+    --seqType fq \
+    --max_memory ${memory}G \
+    --single ${fastq} \
+    --CPU ${task.cpus} \
+    --min_glue ${params.min_glue} \
+    --min_contig_length ${params.min_contig_length} \
+    --output trinity_output_${file_prefix}
+"""
+}
+
+
+process super_transcript {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "$file_id"
+  if (params.assembly_out != "") {
+    publishDir "results/${params.assembly_out}", mode: 'copy'
+  }
+
+  input:
+    tuple val(file_id), path(fasta)
+
+  output:
+    tuple val(file_id), path("trinity_genes.fasta"), path("trinity_genes.gtf"), emit: fasta
+
+  script:
+
+  switch(file_id) {
+    case {it instanceof List}:
+      file_prefix = file_id[0]
+    break;
+    case {it instanceof Map}:
+      file_prefix = file_id.values()[0]
+    break;
+    default:
+      file_prefix = file_id
+    break;
+  };
+  def memory = "${task.memory}" - ~/\s*GB/
+
+"""
+Trinity_gene_splice_modeler.py \
+  --trinity_fasta ${fasta}
+"""
+}

src/nf_modules/ucsc/main.nf

+// SPDX-FileCopyrightText: 2022 Laurent Modolo <laurent.modolo@ens-lyon.fr>
+//
+// SPDX-License-Identifier: AGPL-3.0-or-later
+
+version = "407"
+container_url = "lbmc/ucsc:${version}"
+
+include {
+  index_fasta
+} from './../samtools/main'
+
+params.bedgraph_to_bigwig = ""
+params.bedgraph_to_bigwig_out = ""
+process bedgraph_to_bigwig {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "${file_id}"
+  if (params.bedgraph_to_bigwig_out != "") {
+    publishDir "results/${params.bedgraph_to_bigwig_out}", mode: 'copy'
+  }
+
+  input:
+  tuple val(file_id), path(bg)
+  tuple val(file_id), path(bed)
+
+  output:
+  tuple val(file_id), path("*.bw"), emit: bw
+
+  script:
+"""
+LC_COLLATE=C
+# transform bed file of start-stop chromosome size to stop chromosome size
+awk -v OFS="\\t" '{print \$1, \$3}' ${bed} > chromsize.txt
+
+sort -T ./ -k1,1 -k2,2n ${bg} > \
+  bedGraphToBigWig ${params.bedgraph_to_bigwig} - \
+    chromsize.txt \
+    ${bg.simpleName}_norm.bw
+"""
+}
+
+params.wig_to_bedgraph = ""
+params.wig_to_bedgraph_out = ""
+workflow wig_to_bedgraph {
+  take:
+    fasta
+    wig
+  main:
+    wig_to_bigwig(
+      fasta,
+      wig
+    )
+    bigwig_to_bedgraph(
+      wig_to_bigwig.out.bw
+    )
+  emit:
+  bg = bigwig_to_bedgraph.out.bg
+}
+
+workflow wig2_to_bedgraph2 {
+  take:
+    fasta
+    wig
+  main:
+    wig2_to_bigwig2(
+      fasta,
+      wig
+    )
+    bigwig2_to_bedgraph2(
+      wig2_to_bigwig2.out.bw
+    )
+  emit:
+  bg = bigwig2_to_bedgraph2.out.bg
+}
+
+params.bigwig_to_bedgraph = ""
+params.bigwig_to_bedgraph_out = ""
+process bigwig_to_bedgraph {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "${file_id}"
+  if (params.bigwig_to_bedgraph_out != "") {
+    publishDir "results/${params.bigwig_to_bedgraph_out}", mode: 'copy'
+  }
+
+  input:
+  tuple val(file_id), path(bw)
+
+  output:
+  tuple val(file_id), path("*.bg"), emit: bg
+
+  script:
+"""
+bigWigToBedGraph ${bw} ${bw.simpleName}.bg
+"""
+}
+
+params.bigwig2_to_bedgraph2 = ""
+params.bigwig2_to_bedgraph2_out = ""
+process bigwig2_to_bedgraph2 {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "${file_id}"
+  if (params.bigwig_to_bedgraph_out != "") {
+    publishDir "results/${params.bigwig_to_bedgraph_out}", mode: 'copy'
+  }
+
+  input:
+  tuple val(file_id), path(bw_a), path(bw_b)
+
+  output:
+  tuple val(file_id), path("${bw_a.simpleName}.bg"), path("${bw_b.simpleName}.bg"), emit: bg
+
+  script:
+"""
+bigWigToBedGraph ${bw_a} ${bw_a.simpleName}.bg
+bigWigToBedGraph ${bw_b} ${bw_b.simpleName}.bg
+"""
+}
+
+params.bigwig_to_wig = ""
+params.bigwig_to_wig_out = ""
+process bigwig_to_wig {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "${file_id}"
+  if (params.bigwig_to_wig_out != "") {
+    publishDir "results/${params.bigwig_to_wig_out}", mode: 'copy'
+  }
+
+  input:
+  tuple val(file_id), path(bw)
+
+  output:
+  tuple val(file_id), path("*.wig"), emit: wig
+
+  script:
+"""
+bigWigToBedGraph ${bw} ${bw.simpleName}.bg
+bedgraph_to_wig.pl --bedgraph ${bw.simpleName}.bg --wig ${bw.simpleName}.wig --step 10
+"""
+}
+
+params.bigwig2_to_wig2 = ""
+params.bigwig2_to_wig2_out = ""
+process bigwig2_to_wig2 {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "${file_id}"
+  if (params.bigwig_to_wig_out != "") {
+    publishDir "results/${params.bigwig_to_wig_out}", mode: 'copy'
+  }
+
+  input:
+  tuple val(file_id), path(bw_a), path(bw_b)
+
+  output:
+  tuple val(file_id), path("${bw_a.simpleName}.wig"), path("${bw_b.simpleName}.wig"), emit: wig
+
+  script:
+"""
+bigWigToBedGraph ${bw_a} ${bw_a.simpleName}.bg
+bedgraph_to_wig.pl --bedgraph ${bw_a.simpleName}.bg --wig ${bw_a.simpleName}.wig --step 10
+bigWigToBedGraph ${bw_b} ${bw_b.simpleName}.bg
+bedgraph_to_wig.pl --bedgraph ${bw_b.simpleName}.bg --wig ${bw_b.simpleName}.wig --step 10
+"""
+}
+
+params.wig_to_bigwig = ""
+params.wig_to_bigwig_out = ""
+
+workflow wig_to_bigwig {
+  take:
+    fasta
+    wig
+  main:
+    index_fasta(fasta)
+    wig_to_bigwig_sub(
+      wig,
+      index_fasta.out.index
+    )
+  emit:
+  bw = wig_to_bigwig_sub.out.bw
+}
+
+process wig_to_bigwig_sub {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "${file_id}"
+  if (params.bigwig_to_wig_out != "") {
+    publishDir "results/${params.bigwig_to_wig_out}", mode: 'copy'
+  }
+
+  input:
+  tuple val(file_id), path(w)
+  tuple val(idx_id), path(fasta_idx)
+
+  output:
+  tuple val(file_id), path("${w.simpleName}.bw"), emit: bw
+
+  script:
+"""
+cut -f 1,2 ${fasta_idx} > ${fasta_idx.simpleName}.sizes
+wigToBigWig -clip ${w} ${fasta_idx.simpleName}.sizes ${w.simpleName}.bw
+"""
+}
+
+params.wig2_to_bigwig2 = ""
+params.wig2_to_bigwig2_out = ""
+
+workflow wig2_to_bigwig2 {
+  take:
+    fasta
+    wigs
+  main:
+    index_fasta(fasta)
+    wig2_to_bigwig2_sub(
+      wigs,
+      index_fasta.out.index
+    )
+  emit:
+  bw = wig2_to_bigwig2_sub.out.bw
+}
+
+process wig2_to_bigwig2_sub {
+  container = "${container_url}"
+  label "big_mem_mono_cpus"
+  tag "${file_id}"
+  if (params.bigwig_to_wig_out != "") {
+    publishDir "results/${params.bigwig_to_wig_out}", mode: 'copy'
+  }
+
+  input:
+  tuple val(file_id), path(w_a), path(w_b)
+  tuple val(idx_id), path(fasta_idx)
+
+  output:
+  tuple val(file_id), path("${w_a.simpleName}.bw"), path("${w_b.simpleName}.bw"), emit: bw
+
+  script:
+"""
+cut -f 1,2 ${fasta_idx} > ${fasta_idx.simpleName}.sizes
+wigToBigWig -clip ${w_a} ${fasta_idx.simpleName}.sizes ${w_a.simpleName}.bw
+wigToBigWig -clip ${w_b} ${fasta_idx.simpleName}.sizes ${w_b.simpleName}.bw
+"""
+}
\ No newline at end of file

src/nf_modules/urqt/main.nf

+// SPDX-FileCopyrightText: 2022 Laurent Modolo <laurent.modolo@ens-lyon.fr>
+//
+// SPDX-License-Identifier: AGPL-3.0-or-later
+
+version = "d62c1f8"
+container_url = "lbmc/urqt:${version}"
+
+trim_quality = "20"
+
+params.trimming = "--t 20"
+process trimming {
+  container = "${container_url}"
+  label "big_mem_multi_cpus"
+  tag "${file_id}"
+
+  input:
+  tuple val(file_id), path(reads)
+
+  output:
+  tuple val(pair_id), path("*_trim_R{1,2}.fastq.gz"), emit: fastq
+  path "*_report.txt", emit: report
+
+  script:
+  if (file_id instanceof List){
+    file_prefix = file_id[0]
+  } else {
+    file_prefix = file_id
+  }
+  if (reads.size() == 2)
+"""
+UrQt ${params.trimming} --m ${task.cpus} --gz \
+  --in ${reads[0]} --inpair ${reads[1]} \
+  --out ${file_prefix}_trim_R1.fastq.gz --outpair ${file_prefix}_trim_R2.fastq.gz \
+  > ${pair_id}_trimming_report.txt
+"""
+  else
+"""
+UrQt ${params.trimming} --m ${task.cpus} --gz \
+  --in ${reads[0]} \
+  --out ${file_prefix}_trim.fastq.gz \
+  > ${file_prefix}_trimming_report.txt
+"""
+}
\ No newline at end of file

sge_modules @ 1d6cfb91

-Subproject commit 1d6cfb91449b187b05a3a78df9a06ff3baaf5558

src/solution_RNASeq.nf

+// SPDX-FileCopyrightText: 2022 Laurent Modolo <laurent.modolo@ens-lyon.fr>
+//
+// SPDX-License-Identifier: AGPL-3.0-or-later
+
+nextflow.enable.dsl=2
+
+include { fastp } from "./nf_modules/fastp/main.nf"
+include { fasta_from_bed } from "./nf_modules/bedtools/main.nf"
+include { index_fasta; mapping_fastq } from './nf_modules/kallisto/main.nf' addParams(mapping_fastq_out: "quantification/")
+
+
+params.fastq = "data/fastq/*_{1,2}.fastq"
+
+log.info "fastq files: ${params.fastq}"
+log.info "fasta file : ${params.fasta}"
+log.info "bed file : ${params.bed}"
+
+channel
+  .fromFilePairs( params.fastq, size: -1)
+  .set { fastq_files }
+
+channel
+  .fromPath( params.fasta )
+  .ifEmpty { error "Cannot find any fasta files matching: ${params.fasta}" }
+  .map { it -> [it.simpleName, it]}
+  .set { fasta_files }
+channel
+  .fromPath( params.bed )
+  .ifEmpty { error "Cannot find any bed files matching: ${params.bed}" }
+  .map { it -> [it.simpleName, it]}
+  .set { bed_files }
+
+workflow {
+  fastp(fastq_files)
+  fasta_from_bed(fasta_files, bed_files)
+  index_fasta(fasta_from_bed.out.fasta)
+  mapping_fastq(index_fasta.out.index.collect(), fastp.out.fastq)
+}

src/training_dataset.config

+# SPDX-FileCopyrightText: 2022 Laurent Modolo <laurent.modolo@ens-lyon.fr>
+#
+# SPDX-License-Identifier: AGPL-3.0-or-later
+
 profiles {
   docker {
-    docker.temp = 'auto'
+    docker.temp = "auto"
     docker.enabled = true
     process {
-      $build_synthetic_bed {
-        container = "bedtools:2.25.0"
+      withName: build_synthetic_bed {
+        container = "lbmc/bedtools:2.25.0"
+        cpus = 1
+      }
+      withName: fasta_from_bed {
+        container = "lbmc/bedtools:2.25.0"
+        cpus = 1
+      }
+      withName: index_fasta {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
+      }
+      withName: mapping_fastq_paired {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
+      }
+      withName: bam_2_fastq_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: filter_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: sort_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: index_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: mapping_fastq_single {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
+      }
+      withName: bam_2_fastq_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: filter_bam_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: sort_bam_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: index_bam_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+    }
+  }
+  singularity {
+    singularity.enabled = true
+    singularity.cacheDir = "./bin/"
+    process {
+      withName: build_synthetic_bed {
+        container = "lbmc/bedtools:2.25.0"
+        cpus = 1
+      }
+      withName: fasta_from_bed {
+        container = "lbmc/bedtools:2.25.0"
+        cpus = 1
+      }
+      withName: index_fasta {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
+      }
+      withName: mapping_fastq_single {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
+      }
+      withName: mapping_fastq_paired {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
+      }
+      withName: bam_2_fastq_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: filter_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: sort_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: index_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: bam_2_fastq_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: filter_bam_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: sort_bam_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+      withName: index_bam_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
+      }
+    }
+  }
+  psmn {
+    process{
+      withName: build_synthetic_bed {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/bedtools_2.25.0"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 1
+        memory = "20GB"
+        time = "12h"
+        queue = "monointeldeb128"
       }
-      $fasta_from_bed {
-        container = "bedtools:2.25.0"
+      withName: fasta_from_bed {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/bedtools_2.25.0"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 1
+        memory = "20GB"
+        time = "12h"
+        queue = "monointeldeb128"
       }
-      $index_fasta {
-        container = "bowtie2:2.3.4.1"
+      withName: index_fasta {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/bowtie2_2.3.4.1"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "20GB"
+        time = "12h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $mapping_fastq_paired {
-        container = "bowtie2:2.3.4.1"
+      withName: mapping_fastq_paired {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/bowtie2_2.3.4.1"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $bam_2_fastq_paired {
-        container = "samtools:1.7"
+      withName: bam_2_fastq_paired {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $sort_bam_paired {
-        container = "samtools:1.7"
+      withName: sort_bam_paired {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $index_bam_paired {
-        container = "samtools:1.7"
+      withName: index_bam_paired {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $mapping_fastq_single {
-        container = "bowtie2:2.3.4.1"
+      withName: mapping_fastq_single {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/bowtie2_2.3.4.1"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $bam_2_fastq_single {
-        container = "samtools:1.7"
+      withName: bam_2_fastq_single {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $sort_bam_single {
-        container = "samtools:1.7"
+      withName: sort_bam_single {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      $index_bam_single {
-        container = "samtools:1.7"
+      withName: index_bam_single {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
+        executor = "sge"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 32
+        memory = "30GB"
+        time = "24h"
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
     }
   }
-  sge {
+  ccin2p3 {
+    singularity.enabled = true
+    singularity.cacheDir = "$baseDir/.singularity_in2p3/"
+    singularity.runOptions = "--bind /pbs,/sps,/scratch"
+    process{
+      withName: fasta_from_bed {
+        container = "lbmc/bedtools:2.25.0"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
+      }
+    }
     process{
-      $build_synthetic_bed {
-        beforeScript = "module purge; module load BEDtools/2.25.0"
+      withName: build_synthetic_bed {
+        container = "lbmc/bedtools:2.25.0"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
         executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
         cpus = 1
-        memory = "5GB"
-        time = "6h"
-        queueSize = 1000
-        pollInterval = '60sec'
-        queue = 'h6-E5-2667v4deb128'
-        penv = 'openmp8'
+        queue = "huge"
       }
-      $fasta_from_bed {
-        beforeScript = "module purge; module load BEDtools/2.25.0"
+      withName: fasta_from_bed {
+        container = "lbmc/bedtools:2.25.0"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
         executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
         cpus = 1
-        memory = "5GB"
-        time = "6h"
-        queueSize = 1000
-        pollInterval = '60sec'
-        queue = 'h6-E5-2667v4deb128'
-        penv = 'openmp8'
+        queue = "huge"
       }
-      $index_fasta {
-        beforeScript = "module purge; module load Bowtie2/2.3.4.1"
+      withName: index_fasta {
+        container = "lbmc/bowtie2:2.3.4.1"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $mapping_fastq_paired {
-        beforeScript = "module purge; module load SAMtools/1.7; module load Bowtie2/2.3.4.1"
+      withName: mapping_fastq_paired {
+        container = "lbmc/bowtie2:2.3.4.1"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $bam_2_fastq_paired {
-        beforeScript = "module purge; module load SAMtools/1.7"
+      withName: bam_2_fastq_paired {
+        container = "lbmc/samtools:1.7"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $sort_bam_paired {
-        beforeScript = "module purge; module load SAMtools/1.7"
+      withName: sort_bam_paired {
+        container = "lbmc/samtools:1.7"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $index_bam_paired {
-        beforeScript = "module purge; module load SAMtools/1.7"
+      withName: index_bam_paired {
+        container = "lbmc/samtools:1.7"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $mapping_fastq_single {
-        beforeScript = "module purge; module load SAMtools/1.7; module load Bowtie2/2.3.4.1"
+      withName: mapping_fastq_single {
+        container = "lbmc/bowtie2:2.3.4.1"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $bam_2_fastq_single {
-        beforeScript = "module purge; module load SAMtools/1.7"
+      withName: bam_2_fastq_single {
+        container = "lbmc/samtools:1.7"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $sort_bam_single {
-        beforeScript = "module purge; module load SAMtools/1.7"
+      withName: sort_bam_single {
+        container = "lbmc/samtools:1.7"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
-      $index_bam_single {
-        beforeScript = "module purge; module load SAMtools/1.7"
+      withName: index_bam_single {
+        container = "lbmc/samtools:1.7"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
+        executor = "sge"
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
       }
     }
   }

src/training_dataset.nf

+// SPDX-FileCopyrightText: 2022 Laurent Modolo <laurent.modolo@ens-lyon.fr>
+//
+// SPDX-License-Identifier: AGPL-3.0-or-later
+
 /*
 small pipeline to build a training dataset from whole genome data