Blastn

.editorconfig

@@ -31,3 +31,12 @@ insert_final_newline = unset
 trim_trailing_whitespace = unset
 indent_style = unset
 indent_size = unset
+
+# To prevent errors for these test blastn databases
+[/assets/test*/nt_*/*.{ndb,nhr,nin,nog,nos,not,nsq,ntf,nto}]
+charset = unset
+end_of_line = unset
+insert_final_newline = unset
+trim_trailing_whitespace = unset
+indent_style = unset
+indent_size = unset

.github/CONTRIBUTING.md

@@ -110,4 +110,3 @@ To get started:
 Devcontainer specs:
 
 - [DevContainer config](.devcontainer/devcontainer.json)
-- [Dockerfile](.devcontainer/Dockerfile)

.nf-core.yml

@@ -16,3 +16,5 @@ lint:
   multiqc_config:
     - report_comment
   actions_ci: false
+  template_strings: false
+  merge_markers: false

bin/nohitlist.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+# input
+fasta=$1
+blast=$2
+prefix=$3
+E=$4
+
+# find ids of sequences with no hits in the blastx search
+grep '>' $fasta | \
+    grep -v -w -f <(awk -v evalue="$E" '{{if($14<{evalue}){{print $1}}}}' $blast | sort | uniq) | \
+    cut -f1 | sed 's/>//' > $prefix.nohit.txt
+
+
+

conf/modules.config

@@ -37,12 +37,16 @@ process {
         ext.args = "--evalue 1.0e-25  --max-target-seqs 10 --max-hsps 1"
     }
 
+    withName: "DIAMOND_BLASTX" {
+        ext.args = "--evalue 1.0e-25  --max-target-seqs 10 --max-hsps 1"
+    }
+
     withName: "BLOBTOOLKIT_WINDOWSTATS" {
         ext.args = "--window 0.1 --window 0.01 --window 1 --window 100000 --window 1000000"
     }
 
     withName: "BLOBTOOLKIT_BLOBDIR" {
-        ext.args = "--evalue 1.0e-25 --hit-count 10"
+        ext.args = "--evalue 1.0e-25 --hit-count 10 --update-plot"
         publishDir = [
             path: { "${params.outdir}/" },
             mode: params.publish_dir_mode,
@@ -66,6 +70,26 @@ process {
         ]
     }
 
+    withName: "BLOBTOOLKIT_CHUNK" {
+        ext.args = "--chunk 100000 --overlap 0 --max-chunks 10 --min-length 1000"
+    }
+
+    withName: "BLOBTOOLKIT_UNCHUNK" {
+        ext.args = "--count 10"
+    }
+
+    withName: "NOHIT_LIST" {
+        ext.args = "1.0e-25"
+    }
+
+    withName: "BLAST_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: "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}/blobtoolkit_info" },

conf/test.config

@@ -30,7 +30,9 @@ params {
     taxon     = "Meles meles"
 
     // Databases
-    taxdump   = "/lustre/scratch123/tol/teams/grit/geval_pipeline/btk_databases/taxdump"
-    busco     = "/lustre/scratch123/tol/resources/nextflow/busco_2021_06_reduced/"
-    uniprot   = "${projectDir}/assets/test/mCerEla1.1.buscogenes.dmnd"
+    taxdump = "/lustre/scratch123/tol/teams/grit/geval_pipeline/btk_databases/taxdump"
+    busco   = "/lustre/scratch123/tol/resources/nextflow/busco_2021_06_reduced/"
+    blastp  = "${projectDir}/assets/test/mMelMel3.1.buscogenes.dmnd"
+    blastx  = "${projectDir}/assets/test/mMelMel3.1.buscoregions.dmnd"
+    blastn  = "${projectDir}/assets/test/nt_mMelMel3.1"
 }

conf/test_full.config

@@ -27,7 +27,9 @@ params {
     taxon     = "Laetiporus sulphureus"
 
     // Databases
-    taxdump   = "/lustre/scratch123/tol/teams/grit/geval_pipeline/btk_databases/taxdump"
-    busco     = "/lustre/scratch123/tol/resources/busco/v5/"
-    uniprot   = "${projectDir}/assets/test_full/gfLaeSulp1.1.buscogenes.dmnd"
+    taxdump = "/lustre/scratch123/tol/teams/grit/geval_pipeline/btk_databases/taxdump"
+    busco   = "/lustre/scratch123/tol/resources/busco/v5/"
+    blastp  = "${projectDir}/assets/test_full/gfLaeSulp1.1.buscogenes.dmnd"
+    blastx  = "${projectDir}/assets/test_full/gfLaeSulp1.1.buscoregions.dmnd"
+    blastn  = "${projectDir}/assets/test_full/nt_gfLaeSulp1.1"
 }

modules.json

@@ -5,6 +5,11 @@
         "https://github.com/nf-core/modules.git": {
             "modules": {
                 "nf-core": {
+                    "blast/blastn": {
+                        "branch": "master",
+                        "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
+                        "installed_by": ["modules"]
+                    },
                     "busco": {
                         "branch": "master",
                         "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
@@ -21,6 +26,11 @@
                         "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
                         "installed_by": ["modules"]
                     },
+                    "diamond/blastx": {
+                        "branch": "master",
+                        "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
+                        "installed_by": ["modules"]
+                    },
                     "fastawindows": {
                         "branch": "master",
                         "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
@@ -50,6 +60,11 @@
                         "branch": "master",
                         "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
                         "installed_by": ["modules"]
+                    },
+                    "seqtk/subseq": {
+                        "branch": "master",
+                        "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
+                        "installed_by": ["modules"]
                     }
                 }
             },

modules/local/blastn.nf

@@ -0,0 +1,40 @@
+process BLASTN {
+    tag "$meta.id"
+    label 'process_medium'
+
+    conda "bioconda::blast=2.13.0"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/blast:2.13.0--hf3cf87c_0' :
+        'biocontainers/blast:2.13.0--hf3cf87c_0' }"
+
+    input:
+    tuple val(meta), path(fasta)
+    path  db
+    val   taxid
+
+    output:
+    tuple val(meta), file('*.blastn.txt'), emit: txt
+    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}"
+    def exclude_taxon = taxid ? "-negative_taxids ${taxid}" : ''
+    """
+    DB=`find -L ./ -name "*.ndb" | sed 's/\\.ndb\$//'`
+    blastn \\
+        -num_threads $task.cpus \\
+        -db \$DB \\
+        -query $fasta \\
+        $exclude_taxon \\
+        $args \\
+        -out ${prefix}.blastn.txt
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        blast: \$(blastn -version 2>&1 | sed 's/^.*blastn: //; s/ .*\$//')
+    END_VERSIONS
+    """
+}

modules/local/blobtoolkit/blobdir.nf

@@ -11,7 +11,9 @@ process BLOBTOOLKIT_BLOBDIR {
     tuple val(meta), path(window, stageAs: 'windowstats/*')
     tuple val(meta1), path(busco)
     tuple val(meta2), path(blastp)
-    tuple val(meta3), path(yaml)
+    tuple val(meta3), path(blastx)
+    tuple val(meta4), path(blastn)
+    tuple val(meta5), path(yaml)
     path(taxdump)
 
     output:
@@ -24,15 +26,19 @@ process BLOBTOOLKIT_BLOBDIR {
     script:
     def args = task.ext.args ?: ''
     prefix = task.ext.prefix ?: "${meta.id}"
-    def hits = blastp ? "--hits ${blastp}" : ""
+    def hits_blastp = blastp ? "--hits ${blastp}" : ""
+    def hits_blastx = blastx ? "--hits ${blastx}" : ""
+    def hits_blastn = blastn ? "--hits ${blastn}" : ""
     """
     blobtools replace \\
         --bedtsvdir windowstats \\
         --meta ${yaml} \\
         --taxdump ${taxdump} \\
-        --taxrule buscogenes \\
+        --taxrule bestdistorder=buscoregions \\
         --busco ${busco} \\
-        ${hits} \\
+        ${hits_blastp} \\
+        ${hits_blastx} \\
+        ${hits_blastn} \\
         --threads ${task.cpus} \\
         $args \\
         ${prefix}

modules/local/blobtoolkit/chunk.nf

@@ -0,0 +1,37 @@
+process BLOBTOOLKIT_CHUNK {
+    tag "$meta.id"
+    label 'process_single'
+
+    if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
+        exit 1, "BLOBTOOLKIT_CHUNK module does not support Conda. Please use Docker / Singularity / Podman instead."
+    }
+    container "genomehubs/blobtoolkit:4.1.5"
+
+    input:
+    tuple val(meta) , path(fasta)
+    tuple val(meta2), path(busco_table)
+
+    output:
+    tuple val(meta), path("*.chunks.fasta"), emit: chunks
+    path "versions.yml"                    , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def busco = busco_table ? "--busco ${busco_table}" : "--busco None"
+    """
+    btk pipeline chunk-fasta \\
+        --in ${fasta} \\
+        ${busco} \\
+        --out ${prefix}.chunks.fasta \\
+        $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        blobtoolkit: \$(btk --version | cut -d' ' -f2 | sed 's/v//')
+    END_VERSIONS
+    """
+}

modules/local/blobtoolkit/unchunk.nf

@@ -0,0 +1,34 @@
+process BLOBTOOLKIT_UNCHUNK {
+    tag "$meta.id"
+    label 'process_single'
+
+    if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
+        exit 1, "BLOBTOOLKIT_UNCHUNK module does not support Conda. Please use Docker / Singularity / Podman instead."
+    }
+    container "genomehubs/blobtoolkit:4.1.5"
+
+    input:
+    tuple val(meta), path(blast_table)
+
+    output:
+    tuple val(meta), path("*.out"), emit: blast_out
+    path "versions.yml"           , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${blast_table}"
+    """
+    btk pipeline unchunk-blast \\
+        --in ${blast_table} \\
+        --out ${prefix}.out \\
+        $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        blobtoolkit: \$(btk --version | cut -d' ' -f2 | sed 's/v//')
+    END_VERSIONS
+    """
+}

modules/local/nohit_list.nf

@@ -0,0 +1,32 @@
+process NOHIT_LIST {
+    tag "$meta.id"
+    label 'process_single'
+
+    conda "conda-forge::gawk=5.1.0"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+    'https://depot.galaxyproject.org/singularity/gawk:5.1.0' :
+    'quay.io/biocontainers/gawk:5.1.0' }"
+
+    input:
+    tuple val(meta), path(blast)      //path to blast output table in txt format
+    tuple val(meta), path(fasta)      //path to genome fasta file
+
+    output:
+    tuple val(meta), path ('*.nohit.txt') , emit: nohitlist
+    path "versions.yml"                   , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script: // This script is bundled with the pipeline, in sanger-tol/blobtoolkit/bin/
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    nohitlist.sh ${fasta} ${blast} ${prefix} $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        nohit_list: 1.0
+    END_VERSIONS
+    """
+}