src/singularity_modules/urqt/d62c1f8/urqt.def

-Bootstrap: library
-From: debian:9
-
-%labels
-MAINTAINER Laurent Modolo
-
-%post
-URQT_VERSION=d62c1f8
-PACKAGES="git \
-g++ \
-make \
-ca-certificates \
-zlib1g-dev"
-
-apt-get update && \
-apt-get install -y --no-install-recommends ${PACKAGES} && \
-apt-get clean
-
-git clone https://github.com/l-modolo/UrQt.git && \
-cd UrQt && \
-git checkout ${URQT_VERSION} && \
-make && \
-cd .. && \
-mv UrQt/UrQt /usr/bin/ && \
-rm -Rf UrQt
-
-%environment
-export URQT_VERSION=d62c1f8
-
-%runscript
-exec /bin/bash "$@"

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 {
-      withname: 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 = "bedtools:2.25.0"
+      withName: fasta_from_bed {
+        container = "lbmc/bedtools:2.25.0"
+        cpus = 1
       }
-      withname: index_fasta {
-        container = "bowtie2:2.3.4.1"
+      withName: index_fasta {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
       }
-      withname: mapping_fastq_paired {
-        container = "bowtie2:2.3.4.1"
+      withName: mapping_fastq_paired {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
       }
-      withname: bam_2_fastq_paired {
-        container = "samtools:1.7"
+      withName: bam_2_fastq_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
       }
-      withname: sort_bam_paired {
-        container = "samtools:1.7"
+      withName: filter_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
       }
-      withname: index_bam_paired {
-        container = "samtools:1.7"
+      withName: sort_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
       }
-      withname: mapping_fastq_single {
-        container = "bowtie2:2.3.4.1"
+      withName: index_bam_paired {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
       }
-      withname: bam_2_fastq_single {
-        container = "samtools:1.7"
+      withName: mapping_fastq_single {
+        container = "lbmc/bowtie2:2.3.4.1"
+        cpus = 4
       }
-      withname: sort_bam_single {
-        container = "samtools:1.7"
+      withName: bam_2_fastq_single {
+        container = "lbmc/samtools:1.7"
+        cpus = 4
       }
-      withname: index_bam_single {
-        container = "samtools:1.7"
+      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 /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "BEDtools/2.25.0"
+      withName: build_synthetic_bed {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/bedtools_2.25.0"
         executor = "sge"
-        clusterOptions = "-m be -cwd -V"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 1
         memory = "20GB"
         time = "12h"
-        queue = 'monointeldeb128'
+        queue = "monointeldeb128"
       }
-      withname: fasta_from_bed {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "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 be -cwd -V"
+        clusterOptions = "-m e -cwd -V"
+        cpus = 1
         memory = "20GB"
         time = "12h"
-        queue = 'monointeldeb128'
+        queue = "monointeldeb128"
       }
-      withname: index_fasta {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "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"
+      }
+      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 = 'E5-2670deb128A,E5-2670deb128B,E5-2670deb128C,E5-2670deb128D,E5-2670deb128E,E5-2670deb128F'
-        penv = 'openmp16'
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      withname: mapping_fastq_paired {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "Bowtie2/2.3.4.1: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 = 'E5-2670deb128A,E5-2670deb128B,E5-2670deb128C,E5-2670deb128D,E5-2670deb128E,E5-2670deb128F'
-        penv = 'openmp16'
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      withname: bam_2_fastq_paired {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "SAMtools/1.7"
+      withName: sort_bam_paired {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
         executor = "sge"
-        memory = "20GB"
-        time = "12h"
-        queue = 'monointeldeb128'
+        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"
       }
-      withname: sort_bam_paired {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "SAMtools/1.7"
+      withName: index_bam_paired {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
         executor = "sge"
-        memory = "20GB"
-        time = "12h"
-        queue = 'monointeldeb128'
+        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"
       }
-      withname: index_bam_paired {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "SAMtools/1.7"
+      withName: mapping_fastq_single {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/bowtie2_2.3.4.1"
         executor = "sge"
-        memory = "20GB"
-        time = "12h"
-        queue = 'monointeldeb128'
+        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"
       }
-      withname: mapping_fastq_single {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "Bowtie2/2.3.4.1: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 = 'E5-2670deb128A,E5-2670deb128B,E5-2670deb128C,E5-2670deb128D,E5-2670deb128E,E5-2670deb128F'
-        penv = 'openmp16'
+        queue = "CLG6242deb384A,CLG6242deb384C,CLG5218deb192A,CLG5218deb192B,CLG5218deb192C,CLG5218deb192D,SLG5118deb96,SLG6142deb384A,SLG6142deb384B,SLG6142deb384C,SLG6142deb384D"
+        penv = "openmp32"
       }
-      withname: bam_2_fastq_single {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "SAMtools/1.7"
+      withName: sort_bam_single {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
         executor = "sge"
-        memory = "20GB"
-        time = "12h"
-        queue = 'monointeldeb128'
+        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"
       }
-      withname: sort_bam_single {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "SAMtools/1.7"
+      withName: index_bam_single {
+        beforeScript = "source $baseDir/.conda_psmn.sh"
+        conda = "$baseDir/.conda_envs/samtools_1.7"
         executor = "sge"
-        memory = "20GB"
-        time = "12h"
-        queue = 'monointeldeb128'
+        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"
+      }
+    }
+  }
+  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"
       }
-      withname: index_bam_single {
-        beforeScript = "source /usr/share/lmod/lmod/init/bash; module use ~/privatemodules"
-        module = "SAMtools/1.7"
+    }
+    process{
+      withName: build_synthetic_bed {
+        container = "lbmc/bedtools:2.25.0"
+        scratch = true
+        stageInMode = "copy"
+        stageOutMode = "rsync"
         executor = "sge"
-        memory = "20GB"
-        time = "12h"
-        queue = 'monointeldeb128'
+        clusterOptions = "-P P_lbmc -l os=cl7 -l sps=1 -r n"
+        cpus = 1
+        queue = "huge"
+      }
+      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"
+      }
+      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"
+      }
+      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"
+      }
+      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"
+      }
+      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"
+      }
+      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"
+      }
+      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"
+      }
+      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"
+      }
+      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"
+      }
+      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