update vep and snpeff modules

CHANGELOG.md

@@ -40,6 +40,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#620](https://github.com/nf-core/sarek/pull/620) - Added checks for sex information when running a CNV tools
 - [#623](https://github.com/nf-core/sarek/pull/623) - Additional checks of data in the input sample sheet.
 - [#629](https://github.com/nf-core/sarek/pull/629) - Added checks to catch inconsistency between supplied samples and requested tools.
+- [#632](https://github.com/nf-core/sarek/pull/632) - Added params `--snpeff_version` to allow more configuration on the snpeff container definition
+- [#632](https://github.com/nf-core/sarek/pull/632) - Added params `--vep_include_fasta` to use the fasta file for annotation
 
 ### Changed
 
@@ -81,6 +83,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#608](https://github.com/nf-core/sarek/pull/608) - Prevent candidate VCFs from getting published in manta
 - [#620](https://github.com/nf-core/sarek/pull/620) - `gender` is now `sex` in the samplesheet
 - [#630](https://github.com/nf-core/sarek/pull/630) - Update citations file
+- [#632](https://github.com/nf-core/sarek/pull/632) - Update `snpEff` version to `5.1` and cache up to `105`
+- [#632](https://github.com/nf-core/sarek/pull/632) - Update `VEP` version to `106.1` and cache up to `106`
 
 ### Fixed
 

conf/igenomes.config

@@ -30,12 +30,13 @@ params {
             known_indels          = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/{1000G_phase1,Mills_and_1000G_gold_standard}.indels.b37.vcf"
             known_indels_tbi      = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/{1000G_phase1,Mills_and_1000G_gold_standard}.indels.b37.vcf.idx"
             mappability           = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh37/Annotation/Control-FREEC/out100m2_hg19.gem"
-            snpeff_db             = 'GRCh37.75'
+            snpeff_db             = 'GRCh37.87'
             snpeff_genome         = 'GRCh37'
-            vep_cache_version     = 105
+            snpeff_version        = '5.1'
+            vep_cache_version     = 106
             vep_genome            = 'GRCh37'
             vep_species           = 'homo_sapiens'
-            vep_version           = '104.3'
+            vep_version           = '106.1'
         }
         'GATK.GRCh38' {
             ascat_alleles         = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh38/Annotation/ASCAT/G1000_alleles_hg38.zip"
@@ -60,21 +61,36 @@ params {
             mappability           = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh38/Annotation/Control-FREEC/out100m2_hg38.gem"
             pon                   = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh38/Annotation/GATKBundle/1000g_pon.hg38.vcf.gz"
             pon_tbi               = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh38/Annotation/GATKBundle/1000g_pon.hg38.vcf.gz.tbi"
-            snpeff_db             = 'GRCh38.99'
+            snpeff_db             = 'GRCh38.105'
             snpeff_genome         = 'GRCh38'
-            vep_cache_version     = 105
+            snpeff_version        = '5.1'
+            vep_cache_version     = 106
             vep_genome            = 'GRCh38'
             vep_species           = 'homo_sapiens'
-            vep_version           = '104.3'
+            vep_version           = '106.1'
         }
         'Ensembl.GRCh37' {
             bwa                   = "${params.igenomes_base}/Homo_sapiens/Ensembl/GRCh37/Sequence/BWAIndex/version0.6.0/"
             fasta                 = "${params.igenomes_base}/Homo_sapiens/Ensembl/GRCh37/Sequence/WholeGenomeFasta/genome.fa"
             readme                = "${params.igenomes_base}/Homo_sapiens/Ensembl/GRCh37/Annotation/README.txt"
+            snpeff_db             = 'GRCh37.87'
+            snpeff_genome         = 'GRCh37'
+            snpeff_version        = '5.1'
+            vep_cache_version     = 106
+            vep_genome            = 'GRCh37'
+            vep_species           = 'homo_sapiens'
+            vep_version           = '106.1'
         }
         'NCBI.GRCh38' {
             bwa                   = "${params.igenomes_base}/Homo_sapiens/NCBI/GRCh38/Sequence/BWAIndex/version0.6.0/"
             fasta                 = "${params.igenomes_base}/Homo_sapiens/NCBI/GRCh38/Sequence/WholeGenomeFasta/genome.fa"
+            snpeff_db             = 'GRCh38.105'
+            snpeff_genome         = 'GRCh38'
+            snpeff_version        = '5.1'
+            vep_cache_version     = 106
+            vep_genome            = 'GRCh38'
+            vep_species           = 'homo_sapiens'
+            vep_version           = '106.1'
         }
         'GRCm38' {
             bwa                   = "${params.igenomes_base}/Mus_musculus/Ensembl/GRCm38/Sequence/BWAIndex/version0.6.0/"
@@ -91,10 +107,11 @@ params {
             readme                = "${params.igenomes_base}/Mus_musculus/Ensembl/GRCm38/Annotation/README.txt"
             snpeff_db             = 'GRCm38.99'
             snpeff_genome         = 'GRCm38'
+            snpeff_version        = '5.1'
             vep_cache_version     = 102
             vep_genome            = 'GRCm38'
             vep_species           = 'mus_musculus'
-            vep_version           = '104.3'
+            vep_version           = '106.1'
         }
         'TAIR10' {
             bwa                   = "${params.igenomes_base}/Arabidopsis_thaliana/Ensembl/TAIR10/Sequence/BWAIndex/version0.6.0/"
@@ -114,17 +131,25 @@ params {
         'WBcel235' {
             bwa                   = "${params.igenomes_base}/Caenorhabditis_elegans/Ensembl/WBcel235/Sequence/BWAIndex/version0.6.0/"
             fasta                 = "${params.igenomes_base}/Caenorhabditis_elegans/Ensembl/WBcel235/Sequence/WholeGenomeFasta/genome.fa"
-            snpeff_db             = 'WBcel235.99'
+            snpeff_db             = 'WBcel235.105'
             snpeff_genome         = 'WBcel235'
-            vep_cache_version     = 105
+            snpeff_version        = '5.1'
+            vep_cache_version     = 106
             vep_genome            = 'WBcel235'
             vep_species           = 'caenorhabditis_elegans'
-            vep_version           = '104.3'
+            vep_version           = '106.1'
         }
         'CanFam3.1' {
             bwa                   = "${params.igenomes_base}/Canis_familiaris/Ensembl/CanFam3.1/Sequence/BWAIndex/version0.6.0/"
             fasta                 = "${params.igenomes_base}/Canis_familiaris/Ensembl/CanFam3.1/Sequence/WholeGenomeFasta/genome.fa"
             readme                = "${params.igenomes_base}/Canis_familiaris/Ensembl/CanFam3.1/Annotation/README.txt"
+            snpeff_db             = 'CanFam3.1.99'
+            snpeff_genome         = 'CanFam3.1'
+            snpeff_version        = '5.1'
+            vep_cache_version     = 104
+            vep_genome            = 'CanFam3.1'
+            vep_species           = 'canis_lupus_familiaris'
+            vep_version           = '106.1'
         }
         'GRCz10' {
             bwa                   = "${params.igenomes_base}/Danio_rerio/Ensembl/GRCz10/Sequence/BWAIndex/version0.6.0/"

conf/modules.config

@@ -1133,7 +1133,7 @@ process{
         withName: 'SNPEFF' {
             ext.prefix       = { "${vcf.baseName.minus(".vcf")}_snpEff" }
             ext.args          = '-nodownload -canon -v'
-            if (!params.snpeff_cache) container = { params.snpeff_genome ? "nfcore/snpeff:5.0.${params.snpeff_genome}" : "nfcore/snpeff:5.0.${params.genome}" }
+            if (!params.snpeff_cache) container = { params.snpeff_genome ? "nfcore/snpeff:${params.snpeff_version}.${params.snpeff_genome}" : "nfcore/snpeff:${params.snpeff_version}.${params.genome}" }
             publishDir       = [
                 mode: params.publish_dir_mode,
                 path: { "${params.outdir}/reports/SnpEff/${meta.id}/${meta.variantcaller}" },

conf/test.config

@@ -36,12 +36,13 @@ params {
     germline_resource = params.test_data['homo_sapiens']['genome']['gnomad_r2_1_1_vcf_gz']
     intervals         = params.test_data['homo_sapiens']['genome']['genome_interval_list']
     known_indels      = params.test_data['homo_sapiens']['genome']['mills_and_1000g_indels_vcf_gz']
-    snpeff_db         = 'WBcel235.99'
+    snpeff_db         = 'WBcel235.105'
     snpeff_genome     = 'WBcel235'
-    vep_cache_version = 104
+    snpeff_version    = '5.1'
+    vep_cache_version = 106
     vep_genome        = 'WBcel235'
     vep_species       = 'caenorhabditis_elegans'
-    vep_version       = '104.3'
+    vep_version       = '106.1'
 
     // Ignore params that will throw warning through params validation
     schema_ignore_params = "genomes,test_data"

main.nf

@@ -51,10 +51,11 @@ params.pon                   = WorkflowMain.getGenomeAttribute(params, 'pon')
 params.pon_tbi               = WorkflowMain.getGenomeAttribute(params, 'pon_tbi')
 params.snpeff_db             = WorkflowMain.getGenomeAttribute(params, 'snpeff_db')
 params.snpeff_genome         = WorkflowMain.getGenomeAttribute(params, 'snpeff_genome')
+params.snpeff_version        = WorkflowMain.getGenomeAttribute(params, 'snpeff_version')
 params.vep_cache_version     = WorkflowMain.getGenomeAttribute(params, 'vep_cache_version')
-params.vep_version           = WorkflowMain.getGenomeAttribute(params, 'vep_version')
 params.vep_genome            = WorkflowMain.getGenomeAttribute(params, 'vep_genome')
 params.vep_species           = WorkflowMain.getGenomeAttribute(params, 'vep_species')
+params.vep_version           = WorkflowMain.getGenomeAttribute(params, 'vep_version')
 
 /*
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

modules.json

@@ -67,7 +67,7 @@
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
             "ensemblvep": {
-                "git_sha": "757c20c8ebfab85e1de9f98773a2269f6d34025a"
+                "git_sha": "973151e9eab9bac400aa99f099075a10cdd8e84c"
             },
             "fastp": {
                 "git_sha": "9b51362a532a14665f513cf987531f9ea5046b74"
@@ -214,7 +214,7 @@
                 "git_sha": "6b64f9cb6c3dd3577931cc3cd032d6fb730000ce"
             },
             "snpeff": {
-                "git_sha": "40dd662fd26c3eb3160b7c8cbbe9bff80bbe2c30"
+                "git_sha": "d8bef6057bc2cf657d33ed496426be8416edd780"
             },
             "strelka/germline": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"

nextflow.config

@@ -88,6 +88,7 @@ params {
     annotation_cache = false // Annotation cache disabled
     snpeff_cache = null // No directory for snpEff cache
     vep_cache = null // No directory for VEP cache
+    vep_include_fasta = false // Don't use fasta file for annotation with VEP
 
     // MultiQC options
     multiqc_config             = null