Merge remote-tracking branch 'origin/feat_nftemplate' into docker

CCBR · Jan 12, 2024 · b982a89 · b982a89
2 parents dab67aa + dfce469
commit b982a89
Show file tree

Hide file tree

Showing 18 changed files with 572 additions and 483 deletions.
diff --git a/bin/freec_paired.pl → bin/make_freec_genome_paired.pl b/bin/freec_paired.pl → bin/make_freec_genome_paired.pl
@@ -26,18 +26,19 @@
 print C "chrFiles = $chrFiles\n";
 print C "minimalSubclonePresence = 20\nmaxThreads = 8\n";
 print C "outputDir = $ARGV[0]\n\n";
- 
+
 print C '[sample]' . "\n\n";
- 
+
 print C "mateFile = $tumormateFile\n";
 print C "inputFormat = BAM\nmateOrientation = FR\n\n";
 
-print C '[BAF]' . "\n\n";
+print C '[control]' . "\n\n";
 
 print C "mateFile = $controlmateFile\n";
 print C "inputFormat = BAM\nmateOrientation = FR\n\n";
-
+
+print C '[BAF]' . "\n\n";
 print C "makePileup = $makePileup\n";
 print C "fastaFile = $fastaFile\n";
 print C "minimalCoveragePerPosition = 20\nminimalQualityPerPosition = 20\n";
-print C "SNPfile = $SNPfile";
+print C "SNPfile = $SNPfile";
diff --git a/conf/base.config b/conf/base.config
@@ -51,11 +51,27 @@ process {
     withLabel:process_high_memory {
         memory = { check_max( 200.GB * task.attempt, 'memory' ) }
     }
+    withLabel:process_somaticcaller {
+        cpus   = { check_max( 4     * task.attempt, 'cpus'    ) }
+        memory = { check_max( 64.GB * task.attempt, 'memory'  ) }
+        time   = { check_max( 72.h   * task.attempt, 'time'    ) }
+    }
+    withLabel:process_somaticcaller_high {
+        cpus   = { check_max( 16     * task.attempt, 'cpus'    ) }
+        memory = { check_max( 70.GB * task.attempt, 'memory'  ) }
+        time   = { check_max( 48.h   * task.attempt, 'time'    ) }
+    }
+    withLabel:process_highmem {
+        cpus   = { check_max( 4     * task.attempt, 'cpus'    ) }
+        memory = { check_max( 48.GB * task.attempt, 'memory'  ) }
+        time   = { check_max( 72.h   * task.attempt, 'time'    ) }
+    }
     withLabel:error_ignore {
         errorStrategy = 'ignore'
     }
     withLabel:error_retry {
         errorStrategy = 'retry'
         maxRetries    = 2
     }
+
 }
diff --git a/conf/biowulf.config b/conf/biowulf.config
@@ -22,7 +22,7 @@ singularity {
     autoMounts = true
     cacheDir = "/data/CCBR_Pipeliner/SIFS"
     envWhitelist = 'https_proxy,http_proxy,ftp_proxy,DISPLAY,SLURM_JOBID,SINGULARITY_BINDPATH'
-    runOptions = '-B /gs10,/gs11,/gs12,/gs9,/spin1,/data/CCBR_Pipeliner/,/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/,/data/nousomedr/,/data/CCBR/projects/,/vf/users,/gpfs,/fdb'
+    runOptions = '-B /gs10,/gs11,/gs12,/gs9,/spin1,/data/CCBR_Pipeliner/,/data/CCBR/projects/,/vf/users,/gpfs,/fdb'
 }
 
 env.SINGULARITY_CACHEDIR = "/data/CCBR_Pipeliner/SIFS"
@@ -34,4 +34,4 @@ process {
     stageOutMode = 'rsync'
     // for running pipeline on group sharing data directory, this can avoid inconsistent files timestamps
     cache = 'lenient'
-}
+}
diff --git a/conf/containers.config b/conf/containers.config
@@ -2,6 +2,9 @@
 params {
     containers {
         base = 'nciccbr/ccbr_ubuntu_base_20.04:v6.1'
-        logan = 'docker://dnousome/ccbr_logan_base:v0.3.0' 
+        logan = 'docker://dnousome/ccbr_logan_base:v0.3.3'
+        vcf2maf = 'docker://dnousome/ccbr_vcf2maf:v102.0.0'
+        octopus = 'docker://dancooke/octopus:latest'
+
     }
 }
diff --git a/conf/genomes.config b/conf/genomes.config
@@ -1,11 +1,11 @@
 params {
     genomes {
         'hg38' {
-            genome = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/bwamem2/Homo_sapiens_assembly38.fasta" 
-            genomefai = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/bwamem2/Homo_sapiens_assembly38.fasta.fai" 
+            genome = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/bwamem2/Homo_sapiens_assembly38.fasta"
+            genomefai = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/bwamem2/Homo_sapiens_assembly38.fasta.fai"
             bwagenome= "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/genome/Homo_sapiens_assembly38.fasta"
             genomedict= "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/genome/Homo_sapiens_assembly38.dict"
-            wgsregion = "/data/nousomedr/annotation/resources_broad_hg38_v0_wgs_calling_regions.hg38.interval_list" 
+            wgsregion = "/data/CCBR_Pipeliner/Pipelines/LOGAN/resources/hg38/resources_broad_hg38_v0_wgs_calling_regions.hg38.interval_list"
             intervals= "${projectDir}/assets/hg38_v0_wgs_calling_regions.hg38.bed"
             //millsindel = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/GATK_resource_bundle/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz" //Mills_and_1000G_gold_standard.indels.hg38.vcf.gz"
             //shapeitindel =  "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/GATK_resource_bundle/ALL.wgs.1000G_phase3.GRCh38.ncbi_remapper.20150424.shapeit2_indels.vcf.gz" //ALL.wgs.1000G_phase3.GRCh38.ncbi_remapper.20150424.shapeit2_indels.vcf.gz" //file(params.gold_indels2) //
@@ -14,7 +14,7 @@ params {
             dbsnp = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/GATK_resource_bundle/dbsnp_138.hg38.vcf.gz"
             dbsnp_indel = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/GATK_resource_bundle/dbsnp_indel.vcf"
             gnomad = '--germline-resource /data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/GNOMAD/somatic-hg38-af-only-gnomad.hg38.vcf.gz' // /data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/GNOMAD/somatic-hg38-af-only-gnomad.hg38.vcf.gz
-            pon = "/data/nousomedr/wgs/updatedpon.vcf.gz"    //pon="/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/PON/hg38.noCOSMIC_ClinVar.pon.vcf.gz" //file{params.pon} 
+            pon = "/data/CCBR_Pipeliner/Pipelines/LOGAN/resources/hg38/PON/updatedpon.vcf.gz"    //pon="/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/PON/hg38.noCOSMIC_ClinVar.pon.vcf.gz" //file{params.pon}
             kgp = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/GATK_resource_bundle/1000G_phase1.snps.high_confidence.hg38.vcf.gz"
             KRAKENBACDB = "/data/CCBR_Pipeliner/CCBR_Pipeliner_Legacy/Exome-seek/hg38/kraken/20180907_standard_kraken2"
             snpeff_genome = "GRCh38.86"
@@ -29,11 +29,11 @@ params {
             octopus_gforest= "--forest /data/CCBR_Pipeliner/Pipelines/LOGAN/resources/hg38/octopus/germline.v0.7.4.forest"
             SEQUENZAGC = "/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/hg38/SEQUENZA/hg38_gc50Base.txt.gz"
             chromosomes = ['chr1','chr2','chr3','chr4','chr5','chr6','chr7','chr8','chr9','chr10','chr11','chr12','chr13','chr14','chr15','chr16','chr17','chr18','chr19','chr20','chr21','chr22','chrX','chrY','chrM']
-        }    
+        }
 
         'mm10' {
-            genome = "/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/mm10/genome/bwamem2index/genome.fa" // file(params.genome) 
-            genomefai = "/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/mm10/genome/bwamem2index/genome.fa.fai" // file(params.genome) 
+            genome = "/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/mm10/genome/bwamem2index/genome.fa" // file(params.genome)
+            genomefai = "/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/mm10/genome/bwamem2index/genome.fa.fai" // file(params.genome)
             bwagenome= "/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/mm10/genome/bwaindex/genome.fa"
             genomedict= "/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/mm10/genome/bwamem2index/genome.dict"
             intervals="/data/CCBR_Pipeliner/Pipelines/XAVIER/resources/mm10/genome/bwamem2index/mm10_wgsregions.bed"
@@ -66,6 +66,3 @@ params {
         }
     }
 }
-
-
-
diff --git a/conf/interactive.config b/conf/interactive.config
@@ -4,7 +4,7 @@ params {
     max_cpus = 56
     max_time = '12 h'
 
-    
+
 }
 process.scratch = false
 
@@ -14,4 +14,6 @@ singularity {
     autoMounts = true
     cacheDir = "/data/CCBR_Pipeliner/SIFS"
     envWhitelist='https_proxy,http_proxy,ftp_proxy,DISPLAY,SLURM_JOBID,SINGULARITY_BINDPATH'
-}
+}
+
+env.SINGULARITY_CACHEDIR = "/data/CCBR_Pipeliner/SIFS"
diff --git a/conf/modules.config b/conf/modules.config
@@ -9,27 +9,22 @@ process {
     errorStrategy = 'finish'
 
     withName:'sequenza' {
-        container = 'dnousome/ccbr_logan_base:v0.3.0'
         publishDir = [
             path: { "${params.outdir}/cnv/sequenza" },
             mode: 'copy'
         ]
     }
 
-
     withName: 'freec_paired' {
         publishDir = [
-            path: { "${params.outdir}/cnv/freec" },
+            path: { "${params.outdir}/cnv/freec_paired" },
             mode: 'copy'
         ]
-        container = 'dnousome/ccbr_logan_base:v0.3.0' 
-
     }
 
-    withName:'freec' {
-        container = 'dnousome/ccbr_logan_base:v0.3.0' 
+    withName: 'freec' {
         publishDir = [
-            path: { "${params.outdir}/cnv/freec" },
+            path: { "${params.outdir}/cnv/freec_unpaired" },
             mode: 'copy'
         ]
     }
@@ -192,7 +187,7 @@ process {
             mode: 'copy'
         ]
     }
-      
+
     withName: 'annotsv_tn' {
         publishDir = [
             path: { "${params.outdir}/SV/annotated" },
@@ -291,5 +286,6 @@ process {
             path: { "${params.outdir}/vcfs/combined" },
             mode: 'copy'
         ]
-    }   
+    }
+
 }
diff --git a/main.nf b/main.nf
@@ -5,14 +5,12 @@ date = new Date().format( 'yyyyMMdd' )
 
 
 log.info """\
-         L O G A N     P I P E L I N E    
+         L O G A N     P I P E L I N E
          =============================
          genome: ${params.genome}
          outdir: ${params.outdir}
          Sample Sheet: ${params.sample_sheet}
-         Samples: ${params.fastq_input} ${params.file_input} ${params.bam_input}
-         NF version   : $nextflow.version
-
+         Samples: ${params.fastq_input} ${params.fastq_file_input} ${params.bam_input} ${params.bam_file_input}
          """
          .stripIndent()
 
@@ -40,8 +38,7 @@ workflow.onComplete {
 
 //Final Workflow
 workflow {
-    DETERMINEBAM()
-    if ([params.fastq_input,params.file_input].any() && params.sample_sheet && !params.BAMINPUT){
+        if ([params.fastq_input,params.fastq_file_input].any() && params.sample_sheet){
         println "Tumor-Normal FASTQ"
         INPUT()
         ALIGN(INPUT.out.fastqinput,INPUT.out.sample_sheet)
@@ -52,10 +49,10 @@ workflow {
         //Tumor-Normal VC, SV, CNV
         if (params.vc){
             VC(ALIGN.out.bamwithsample,ALIGN.out.splitout,ALIGN.out.sample_sheet)
-        }   
+        }
         if (params.sv){
             SV(ALIGN.out.bamwithsample)
-        }  
+        }
         if (params.cnv){
             if (params.genome == "mm10"){
                 CNVmouse(ALIGN.out.bamwithsample)
@@ -67,41 +64,41 @@ workflow {
                     CNVhuman(ALIGN.out.bamwithsample,VC.out.somaticcall_input)
                 }
             }
-    }  
+    }
         if (params.qc && params.gl){
             QC_GL(ALIGN.out.fastqin,ALIGN.out.fastpout,ALIGN.out.bqsrout,GL.out.glnexusout,GL.out.bcfout)
         }  else if (params.qc){
             QC_NOGL(ALIGN.out.fastqin,ALIGN.out.fastpout,ALIGN.out.bqsrout)
-        }  
+        }
 
     }
-    
+
     //TUMOR-NOMRAL BAM INPUT
-    if ([params.bam_input,params.file_input].any() && params.sample_sheet && BAMINPUT){
+    if ([params.bam_input,params.bam_file_input].any() && params.sample_sheet){
         println "Tumor-Normal with BAMs"
         INPUT_BAM()
         if (params.vc){
             VC(INPUT_BAM.out.bamwithsample,INPUT_BAM.out.splitout,INPUT_BAM.out.sample_sheet)
-        }  
+        }
         if (params.sv){
             SV(INPUT_BAM.out.bamwithsample)
-        }  
+        }
         if (params.cnv){
             if (params.genome == "mm10"){
                 CNVmouse(INPUT_BAM.out.bamwithsample)
             } else if (params.genome== "hg38"){
                 if (!params.vc){
                     VC(INPUT_BAM.out.bamwithsample,INPUT_BAM.out.splitout,INPUT_BAM.out.sample_sheet)
                     CNVhuman(INPUT_BAM.out.bamwithsample,VC.out.somaticcall_input)
-                }else { 
+                }else {
                     CNVhuman(INPUT_BAM.out.bamwithsample,VC.out.somaticcall_input)
                 }
             }
         }
-    }  
-    
+    }
+
     ///Tumor Only Pipelines
-    if ([params.fastq_input,params.file_input].any() && !params.sample_sheet && !params.BAMINPUT){
+    if ([params.fastq_input,params.file_input].any() && !params.sample_sheet){
         println "Tumor-Only FASTQ"
         INPUT_TONLY()
         ALIGN_TONLY(INPUT_TONLY.out.fastqinput,INPUT_TONLY.out.sample_sheet)
@@ -118,7 +115,7 @@ workflow {
                 if (!params.vc){
                     VC_TONLY(ALIGN_TONLY.out.bamwithsample,ALIGN_TONLY.out.splitout,ALIGN_TONLY.out.sample_sheet)
                     CNVhuman_tonly(ALIGN_TONLY.out.bamwithsample,VC_TONLY.out.somaticcall_input)
-                } else{
+                } else {
                     CNVhuman_tonly(ALIGN_TONLY.out.bamwithsample,VC_TONLY.out.somaticcall_input)
                 }
             }
@@ -129,28 +126,27 @@ workflow {
     }
 
     //Variant Calling from BAM-Tumor Only Mode
-    if ([params.bam_input,params.file_input].any() && !params.sample_sheet && params.BAMINPUT){
+    if ([params.bam_input,params.bam_file_input].any() && !params.sample_sheet){
         println "Tumor-Only BAM"
         INPUT_TONLY_BAM()
         if (params.vc){
             VC_TONLY(INPUT_TONLY_BAM.out.bamwithsample,INPUT_TONLY_BAM.out.splitout,INPUT_TONLY_BAM.out.sample_sheet)
         }
         if (params.sv){
             SV_TONLY(INPUT_TONLY_BAM.out.bamwithsample)
-        }  
+        }
         if (params.cnv){
             if (params.genome == "mm10"){
                 CNVmouse_tonly(INPUT_TONLY_BAM.out.bamwithsample)
             } else if (params.genome== "hg38"){
-                VC_TONLY(INPUT_TONLY_BAM.out.bamwithsample,INPUT_TONLY_BAM.out.splitout,INPUT_TONLY_BAM.out.sample_sheet)
-                CNVhuman_tonly(INPUT_TONLY_BAM.out.bamwithsample,VC_TONLY.out.somaticcall_input)
-            }
-        }  
-
-    }
+                if (!params.vc){
+                    VC_TONLY(INPUT_TONLY_BAM.out.bamwithsample,INPUT_TONLY_BAM.out.splitout,INPUT_TONLY_BAM.out.sample_sheet)
+                    CNVhuman_tonly(INPUT_TONLY_BAM.out.bamwithsample,VC_TONLY.out.somaticcall_input)
+                } else {
+                    CNVhuman_tonly(INPUT_TONLY_BAM.out.bamwithsample,VC_TONLY.out.somaticcall_input)
+                }
 
+        }
+        }
+    }
 }
-
-
-
-