get the pipeline working

sanger-tol · Dec 18, 2023 · da3b983 · da3b983
1 parent 3af990f
commit da3b983
Show file tree

Hide file tree

Showing 28 changed files with 183 additions and 523 deletions.
diff --git a/conf/igenomes.config b/conf/igenomes.config
diff --git a/conf/modules.config b/conf/modules.config
@@ -63,7 +63,7 @@ process {
     withName: "BUSCO" {
         scratch = true
         // Overridden in the test profile, see at the end of this file
-        ext.args = "--mode genome --force"
+        ext.args = "--force"
     }
 
     withName: "DIAMOND_BLASTP" {
@@ -74,15 +74,24 @@ process {
         ext.args = "--evalue 1.0e-25  --max-target-seqs 10 --max-hsps 1"
     }
 
+    withName: "BLOBTOOLKIT_DEPTH" {
+        ext.args = "-s 1000"
+    }
+
     withName: "BLOBTOOLKIT_WINDOWSTATS" {
         ext.args = "--window 0.1 --window 0.01 --window 1 --window 100000 --window 1000000"
     }
 
     withName: "BLOBTOOLKIT_CREATEBLOBDIR" {
         ext.args = "--evalue 1.0e-25 --hit-count 10"
+        publishDir = [
+            path: { "${params.outdir}/" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals("versions.yml") ? null : filename }
+        ]
     }
 
-    withName: "BLOBTOOLKIT_CREATEBLOBDIR" {
+    withName: "BLOBTOOLKIT_UPDATEBLOBDIR" {
         ext.args = "--evalue 1.0e-25 --hit-count 10 --update-plot"
         publishDir = [
             path: { "${params.outdir}/" },
@@ -123,10 +132,6 @@ process {
         ext.args = "-outfmt '6 qseqid staxids bitscore std' -max_target_seqs 10 -max_hsps 1 -evalue 1.0e-10 -lcase_masking -dust '20 64 1'"
     }
 
-    withName: "BLASTN" {
-        ext.args = "-outfmt '6 qseqid staxids bitscore std' -max_target_seqs 10 -max_hsps 1 -evalue 1.0e-10 -lcase_masking -dust '20 64 1'"
-    }
-
     withName: "CUSTOM_DUMPSOFTWAREVERSIONS" {
         publishDir = [
             path: { "${params.outdir}/pipeline_info/blobtoolkit" },
@@ -159,7 +164,7 @@ profiles {
         process {
             withName: BUSCO {
                 // Note: BUSCO *must* see the double-quotes around the parameters
-                ext.args = '--mode genome --force --metaeuk_parameters \'"-s=2"\' --metaeuk_rerun_parameters \'"-s=2"\''
+                ext.args = '--force --metaeuk_parameters \'"-s=2"\' --metaeuk_rerun_parameters \'"-s=2"\''
             }
         }
     }

diff --git a/conf/test_raw.config b/conf/test_raw.config
@@ -33,5 +33,7 @@ params {
     // Databases
     taxdump   = "/lustre/scratch123/tol/teams/grit/geval_pipeline/btk_databases/taxdump"
     busco     = "/lustre/scratch123/tol/resources/nextflow/busco/blobtoolkit.GCA_922984935.2.2023-08-03"
-    uniprot   = "https://tolit.cog.sanger.ac.uk/test-data/resources/diamond/mCerEla1.1.buscogenes.dmnd"
+    blastp  = "${projectDir}/assets/test/mMelMel3.1.buscogenes.dmnd"
+    blastx  = "${projectDir}/assets/test/mMelMel3.1.buscoregions.dmnd"
+    blastn  = "${projectDir}/assets/test/nt_mMelMel3.1/"
 }
diff --git a/main.nf b/main.nf
@@ -10,17 +10,6 @@
 
 nextflow.enable.dsl = 2
 
-/*
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-    GENOME PARAMETER VALUES
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-*/
-
-// Remove this line if you don't need a FASTA file
-//   This is an example of how to use getGenomeAttribute() to fetch parameters
-//   from igenomes.config using `--genome`
-// params.fasta = WorkflowMain.getGenomeAttribute(params, 'fasta')
-
 /*
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     VALIDATE & PRINT PARAMETER SUMMARY

diff --git a/modules.json b/modules.json
@@ -10,7 +10,8 @@
                         "git_sha": "f0d13ae7e1f9b24a705764f8673af859268d7077",
                         "installed_by": [
                             "modules"
-                        ]
+                        ],
+                        "patch": "modules/nf-core/blast/blastn/blast-blastn.diff"
                     },
                     "busco": {
                         "branch": "master",

diff --git a/modules/local/blobtoolkit/chunk.nf b/modules/local/blobtoolkit/chunk.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_CHUNK {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_CHUNK module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta) , path(fasta)

diff --git a/modules/local/blobtoolkit/config.nf b/modules/local/blobtoolkit/config.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_CONFIG {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "GENERATE_CONFIG module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(fasta)

diff --git a/modules/local/blobtoolkit/countbuscos.nf b/modules/local/blobtoolkit/countbuscos.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_COUNTBUSCOS {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_COUNTBUSCOS module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(table, stageAs: 'dir??/*')

diff --git a/modules/local/blobtoolkit/createblobdir.nf b/modules/local/blobtoolkit/createblobdir.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_CREATEBLOBDIR {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_BLOBDIR module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(window, stageAs: 'windowstats/*')

diff --git a/modules/local/blobtoolkit/depth.nf b/modules/local/blobtoolkit/depth.nf
@@ -0,0 +1,35 @@
+process BLOBTOOLKIT_DEPTH {
+    tag "${meta.id}"
+    label 'process_single'
+
+    if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
+        exit 1, "BLOBTOOLKIT_DEPTH module does not support Conda. Please use Docker / Singularity / Podman instead."
+    }
+    container "docker.io/genomehubs/blobtk:0.5.1"
+
+    input:
+    tuple val(meta), path(bam), path(bai)
+    each plot
+
+    output:
+    tuple val(meta), path('*.regions.bed.gz') , emit: bed
+    path "versions.yml"                       , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    blobtk depth \\
+        -b ${bam} \\
+        $args \\
+        -O ${prefix}.regions.bed.gz \\
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        blobtoolkit: \$(btk --version | cut -d' ' -f2 | sed 's/v//')
+    END_VERSIONS
+    """
+}
diff --git a/modules/local/blobtoolkit/extractbuscos.nf b/modules/local/blobtoolkit/extractbuscos.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_EXTRACTBUSCOS {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_EXTRACTBUSCOS module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(fasta)

diff --git a/modules/local/blobtoolkit/images.nf b/modules/local/blobtoolkit/images.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_IMAGES {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_IMAGES module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtk:0.3.3"
+    container "docker.io/genomehubs/blobtk:0.5.1"
 
     input:
     tuple val(meta), path(blobdir)

diff --git a/modules/local/blobtoolkit/metadata.nf b/modules/local/blobtoolkit/metadata.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_METADATA {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_METADATA module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(yaml)

diff --git a/modules/local/blobtoolkit/summary.nf b/modules/local/blobtoolkit/summary.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_SUMMARY {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_SUMMARY module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(blobdir)

diff --git a/modules/local/blobtoolkit/unchunk.nf b/modules/local/blobtoolkit/unchunk.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_UNCHUNK {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_UNCHUNK module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(blast_table)

diff --git a/modules/local/blobtoolkit/updateblobdir.nf b/modules/local/blobtoolkit/updateblobdir.nf
@@ -5,11 +5,12 @@ process BLOBTOOLKIT_UPDATEBLOBDIR {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "BLOBTOOLKIT_BLOBDIR module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(input)
-    tuple val(meta1), path(blastx)
+    tuple val(meta1), path(blastx, stageAs: "blastx.txt")
+    tuple val(meta2), path(blastn, stageAs: "blastn.txt")
     path(taxdump)
 
     output:
@@ -23,13 +24,13 @@ process BLOBTOOLKIT_UPDATEBLOBDIR {
     def args = task.ext.args ?: ''
     prefix = task.ext.prefix ?: "${meta.id}"
     def hits_blastx = blastx ? "--hits ${blastx}" : ""
+    def hits_blastn = blastn ? "--hits ${blastn}" : ""
     """
     blobtools replace \\
         --taxdump ${taxdump} \\
         --taxrule bestdistorder=buscoregions \\
         ${hits_blastx} \\
-        --bedtsvdir windowstats \\
-        --meta ${yaml} \\
+        ${hits_blastn} \\
         --threads ${task.cpus} \\
         $args \\
         ${input}

diff --git a/modules/local/blobtoolkit/windowstats.nf b/modules/local/blobtoolkit/windowstats.nf
@@ -5,7 +5,7 @@ process BLOBTOOLKIT_WINDOWSTATS {
     if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
         exit 1, "GET_WINDOW_STATS module does not support Conda. Please use Docker / Singularity / Podman instead."
     }
-    container "docker.io/genomehubs/blobtoolkit:4.1.5"
+    container "docker.io/genomehubs/blobtoolkit:4.3.0"
 
     input:
     tuple val(meta), path(tsv)

diff --git a/modules/nf-core/blast/blastn/blast-blastn.diff b/modules/nf-core/blast/blastn/blast-blastn.diff
diff --git a/modules/nf-core/blast/blastn/main.nf b/modules/nf-core/blast/blastn/main.nf
diff --git a/nextflow.config b/nextflow.config
@@ -201,12 +201,6 @@ plugins {
     id 'nf-validation' // Validation of pipeline parameters and creation of an input channel from a sample sheet
 }
 
-// Load igenomes.config if required
-if (!params.igenomes_ignore) {
-    includeConfig 'conf/igenomes.config'
-} else {
-    params.genomes = [:]
-}
 // Export these variables to prevent local Python/R libraries from conflicting with those in the container
 // The JULIA depot path has been adjusted to a fixed path `/usr/local/share/julia` that needs to be used for packages in the container.
 // See https://apeltzer.github.io/post/03-julia-lang-nextflow/ for details on that. Once we have a common agreement on where to keep Julia packages, this is adjustable.