Merge pull request #386 from maxulysse/dsl2_annotation

Dsl2 annotation
nf-core · Jun 29, 2021 · 88ac395 · 88ac395
2 parents df65564 + b7781ea
commit 88ac395
Show file tree

Hide file tree

Showing 44 changed files with 1,107 additions and 638 deletions.
diff --git a/.circleci/config.yml b/.circleci/config.yml
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -26,7 +26,7 @@ jobs:
         nxf_ver: ['21.04.0', '']
         engine: ['docker']
 
-        test: ['default', 'aligner', 'gatk4_spark', 'targeted', 'tumor_normal_pair', 'variant_calling']
+        test: ['default', 'aligner', 'gatk4_spark', 'targeted', 'tumor_normal_pair', 'variant_calling', 'annotation']
     steps:
       - name: Check out pipeline code
         uses: actions/checkout@v2

diff --git a/conf/base.config b/conf/base.config
@@ -41,19 +41,19 @@ process {
     time = { check_resource( 20.h * task.attempt) }
   }
 
-//  withName:CONCAT_VCF {
-//    // For unknown reasons, CONCAT_VCF sometimes fails with SIGPIPE
-//    // (exit code 141). Rerunning the process will usually work.
-//    errorStrategy = {task.exitStatus == 141 ? 'retry' : 'terminate'}
-//  }
-  withLabel:FASTQC {
+  withName:CONCAT_VCF {
+    // For unknown reasons, CONCAT_VCF sometimes fails with SIGPIPE
+    // (exit code 141). Rerunning the process will usually work.
+    errorStrategy = {task.exitStatus == 141 ? 'retry' : 'terminate'}
+  }
+  withName:FASTQC {
     errorStrategy = {task.exitStatus == 143 ? 'retry' : 'ignore'}
   }
-  withLabel:BWAMEM2_MEM {
+  withName:BWAMEM2_MEM {
     memory = {check_resource(60.GB * task.attempt)}
     time = {check_resource(48.h * task.attempt)}
   }
-//  withName:MULTIQC {
-//    errorStrategy = {task.exitStatus == 143 ? 'retry' : 'ignore'}
-//  }
+  withName:MULTIQC {
+    errorStrategy = {task.exitStatus == 143 ? 'retry' : 'ignore'}
+  }
 }
diff --git a/conf/genomes.config b/conf/genomes.config
@@ -27,8 +27,9 @@ params {
       known_indels_index      = "${params.genomes_base}/{1000G_phase1,Mills_and_1000G_gold_standard}.indels.b37.vcf.idx"
       mappability             = "${params.genomes_base}/out100m2_hg19.gem"
       snpeff_db               = 'GRCh37.75'
-      species                 = 'homo_sapiens'
-      vep_cache_version       = '99'
+      vep_genome              = 'GRCh37'
+      vep_species             = 'homo_sapiens'
+      vep_cache_version       = '104'
     }
     'GRCh38' {
       ac_loci                 = "${params.genomes_base}/1000G_phase3_GRCh38_maf0.3.loci"
@@ -47,9 +48,10 @@ params {
       known_indels            = "${params.genomes_base}/{Mills_and_1000G_gold_standard.indels.hg38,beta/Homo_sapiens_assembly38.known_indels}.vcf.gz"
       known_indels_index      = "${params.genomes_base}/{Mills_and_1000G_gold_standard.indels.hg38,beta/Homo_sapiens_assembly38.known_indels}.vcf.gz.tbi"
       mappability             = "${params.genomes_base}/out100m2_hg38.gem"
-      snpeff_db               = 'GRCh38.86'
-      species                 = 'homo_sapiens'
-      vep_cache_version       = '99'
+      snpeff_db               = 'GRCh38.99'
+      vep_genome              = 'GRCh38'
+      vep_species             = 'homo_sapiens'
+      vep_cache_version       = '104'
     }
     'minimalGRCh37' {
       fasta                   = "${params.genomes_base}/human_g1k_v37_decoy.small.fasta"
@@ -60,9 +62,10 @@ params {
       germline_resource       = "${params.genomes_base}/gnomAD.r2.1.1.GRCh37.small.PASS.AC.AF.only.vcf.gz"
       intervals               = "${params.genomes_base}/small.intervals"
       known_indels            = "${params.genomes_base}/Mills_1000G_gold_standard_and_1000G_phase1.indels.b37.small.vcf.gz"
-      snpeff_db               = 'GRCh37.87'
-      species                 = 'homo_sapiens'
-      vep_cache_version       = '99'
+      snpeff_db               = 'GRCh37.75'
+      vep_genome              = 'GRCh37'
+      vep_species             = 'homo_sapiens'
+      vep_cache_version       = '104'
     }
     'smallerGRCh37' {
       fasta                   = "${params.genomes_base}/human_g1k_v37_decoy.small.fasta"
@@ -72,8 +75,9 @@ params {
       dbsnp                   = "${params.genomes_base}/dbsnp_146_hg38_chr20_tso-only.vcf.gz"
       fasta                   = "${params.genomes_base}/Homo_sapiens_assembly38_chr20.fasta"
       known_indels            = "${params.genomes_base}/Mills_and_1000G_gold_standard_indels_hg38_chr20.vcf.gz"
-      snpeff_db               = 'GRCh38.86'
-      species                 = 'homo_sapiens'
+      snpeff_db               = 'GRCh38.99'
+      vep_genome              = 'GRCh38'
+      vep_species             = 'homo_sapiens'
       vep_cache_version       = '99'
     }
     'custom' {
@@ -85,9 +89,10 @@ params {
       fasta_fai               = "${params.genomes_base}/data/genomics/homo_sapiens/genome/genome.fasta.fai"
       germline_resource       = "${params.genomes_base}/data/genomics/homo_sapiens/genome/vcf/gnomAD.r2.1.1.vcf.gz"
       known_indels            = "${params.genomes_base}/data/genomics/homo_sapiens/genome/vcf/mills_and_1000G.indels.vcf.gz"
-      snpeff_db               = 'GRCh38.86'
-      species                 = 'homo_sapiens'
-      vep_cache_version       = '99'
+      snpeff_db               = 'GRCh38.99'
+      vep_genome              = 'GRCh38'
+      vep_species             = 'homo_sapiens'
+      vep_cache_version       = '104'
     }
   }
 }
diff --git a/conf/igenomes.config b/conf/igenomes.config
@@ -28,8 +28,9 @@ params {
       known_indels_index      = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/{1000G_phase1,Mills_and_1000G_gold_standard}.indels.b37.vcf.idx"
       mappability             = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh37/Annotation/Control-FREEC/out100m2_hg19.gem"
       snpeff_db               = 'GRCh37.75'
-      species                 = 'homo_sapiens'
-      vep_cache_version       = '99'
+      vep_cache_version       = '104'
+      vep_genome              = 'GRCh37'
+      vep_species             = 'homo_sapiens'
     }
     'Ensembl.GRCh37' {
       fasta                   = "${params.igenomes_base}/Homo_sapiens/Ensembl/GRCh37/Sequence/WholeGenomeFasta/genome.fa"
@@ -52,9 +53,10 @@ params {
       known_indels            = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh38/Annotation/GATKBundle/{Mills_and_1000G_gold_standard.indels.hg38,beta/Homo_sapiens_assembly38.known_indels}.vcf.gz"
       known_indels_index      = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh38/Annotation/GATKBundle/{Mills_and_1000G_gold_standard.indels.hg38,beta/Homo_sapiens_assembly38.known_indels}.vcf.gz.tbi"
       mappability             = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh38/Annotation/Control-FREEC/out100m2_hg38.gem"
-      snpeff_db               = 'GRCh38.86'
-      species                 = 'homo_sapiens'
-      vep_cache_version       = '99'
+      snpeff_db               = 'GRCh38.99'
+      vep_cache_version       = '104'
+      vep_genome              = 'GRCh38'
+      vep_species             = 'homo_sapiens'
     }
     'NCBI.GRCh38' {
       fasta                   = "${params.igenomes_base}/Homo_sapiens/NCBI/GRCh38/Sequence/WholeGenomeFasta/genome.fa"
@@ -73,9 +75,10 @@ params {
       known_indels            = "${params.igenomes_base}/Mus_musculus/Ensembl/GRCm38/MouseGenomeProject/mgp.v5.merged.indels.dbSNP142.normed.vcf.gz"
       known_indels_index      = "${params.igenomes_base}/Mus_musculus/Ensembl/GRCm38/MouseGenomeProject/mgp.v5.merged.indels.dbSNP142.normed.vcf.gz.tbi"
       mappability             = "${params.igenomes_base}/Mus_musculus/Ensembl/GRCm38/Annotation/Control-FREEC/GRCm38_68_mm10.gem"
-      snpeff_db               = 'GRCm38.86'
-      species                 = 'mus_musculus'
-      vep_cache_version       = '99'
+      snpeff_db               = 'GRCm38.99'
+      vep_cache_version       = '102'
+      vep_genome              = 'GRCm38'
+      vep_species             = 'mus_musculus'
     }
     'TAIR10' {
       bwa                     = "${params.igenomes_base}/Arabidopsis_thaliana/Ensembl/TAIR10/Sequence/BWAIndex/genome.fa.{amb,ann,bwt,pac,sa}"
@@ -96,9 +99,10 @@ params {
     'WBcel235' {
       bwa                     = "${params.igenomes_base}/Caenorhabditis_elegans/Ensembl/WBcel235/Sequence/BWAIndex/genome.fa.{amb,ann,bwt,pac,sa}"
       fasta                   = "${params.igenomes_base}/Caenorhabditis_elegans/Ensembl/WBcel235/Sequence/WholeGenomeFasta/genome.fa"
-      snpeff_db               = 'WBcel235.86'
-      species                 = 'caenorhabditis_elegans'
-      vep_cache_version       = '99'
+      snpeff_db               = 'WBcel235.99'
+      vep_cache_version       = '104'
+      vep_genome              = 'WBcel235'
+      vep_species             = 'caenorhabditis_elegans'
     }
     'ce10' {
       bwa                     = "${params.igenomes_base}/Caenorhabditis_elegans/UCSC/ce10/Sequence/BWAIndex/genome.fa.{amb,ann,bwt,pac,sa}"

diff --git a/conf/modules.config b/conf/modules.config
@@ -27,7 +27,8 @@ params {
             publish_dir      = 'reference'
             publish_files    = false
         }
-        'index_target_bed' {
+        'bgziptabix_target_bed' {
+            suffix           = '.bed'
             publish_dir      = 'reference'
             publish_files    = false
         }
@@ -200,6 +201,36 @@ params {
             publish_dir      = 'variant_calling'
             publish_files    = ['vcf.gz':'strelka', 'vcf.gz.tbi':'strelka']
         }
+// ANNOTATE
+        'snpeff' {
+            args             = '-nodownload -canon -v'
+            publish_files    = false
+            tag_base         = '5.0'
+        }
+        'bgziptabix_snpeff' {
+            publish_dir      = 'annotation'
+            suffix           = '_snpEff.ann'
+            publish_by_meta  = true
+        }
+        'merge_vep' {
+            args             = '--everything --filter_common --per_gene --total_length --offline'
+            publish_files    = false
+        }
+        'bgziptabix_merge_vep' {
+            publish_dir      = 'annotation'
+            suffix           = '_snpEff_VEP.ann'
+            publish_by_meta  = true
+        }
+        'vep' {
+            args             = '--everything --filter_common --per_gene --total_length --offline'
+            publish_files    = false
+            tag_base         = '104.3'
+        }
+        'bgziptabix_vep' {
+            publish_dir      = 'annotation'
+            suffix           = '_VEP.ann'
+            publish_by_meta  = true
+        }
 // QC_TRIM
         'fastqc' {
             args             = '--quiet'