Split preseq lcextrap and ccurve (#1440)

* fix(preseq): Update lcextrap file name

* chore(preseq): Bump lcextrap image

* feat: Initialize preseq c_curve

* docs(preseq): Update documentation

modules/preseq/ccurve/main.nf

@@ -0,0 +1,40 @@
+process PRESEQ_CCURVE {
+    tag "$meta.id"
+    label 'process_medium'
+    label 'error_ignore'
+
+    conda (params.enable_conda ? "bioconda::preseq=3.1.2" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/preseq:3.1.2--h445547b_2':
+        'quay.io/biocontainers/preseq:3.1.2--h445547b_2' }"
+
+    input:
+    tuple val(meta), path(bam)
+
+    output:
+    tuple val(meta), path("*.c_curve.txt"), emit: c_curve
+    tuple val(meta), path("*.log")        , emit: log
+    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 paired_end = meta.single_end ? '' : '-pe'
+    """
+    preseq \\
+        c_curve \\
+        $args \\
+        $paired_end \\
+        -output ${prefix}.c_curve.txt \\
+        $bam
+    cp .command.err ${prefix}.command.log
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        preseq: \$(echo \$(preseq 2>&1) | sed 's/^.*Version: //; s/Usage:.*\$//')
+    END_VERSIONS
+    """
+}

modules/preseq/ccurve/meta.yml

@@ -0,0 +1,48 @@
+name: preseq_ccurve
+description: Software for predicting library complexity and genome coverage in high-throughput sequencing
+keywords:
+  - preseq
+  - library
+  - complexity
+tools:
+  - preseq:
+      description: Software for predicting library complexity and genome coverage in high-throughput sequencing
+      homepage: http://smithlabresearch.org/software/preseq/
+      documentation: http://smithlabresearch.org/wp-content/uploads/manual.pdf
+      tool_dev_url: https://github.com/smithlabcode/preseq
+      doi: ""
+      licence: ["GPL"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - bam:
+      type: file
+      description: BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - ccurve:
+      type: file
+      description: File containing output of Preseq c curve
+      pattern: "*.{c_curve.txt}"
+  - log:
+      type: file
+      description: Log file containing stderr produced by Preseq
+      pattern: "*.{log}"
+
+authors:
+  - "@drpatelh"
+  - "@Emiller88"

modules/preseq/lcextrap/main.nf

@@ -5,16 +5,16 @@ process PRESEQ_LCEXTRAP {
 
     conda (params.enable_conda ? "bioconda::preseq=3.1.2" : null)
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/preseq:3.1.2--h06ef8b0_1' :
-        'quay.io/biocontainers/preseq:3.1.2--h06ef8b0_1' }"
+        'https://depot.galaxyproject.org/singularity/preseq:3.1.2--h445547b_2':
+        'quay.io/biocontainers/preseq:3.1.2--h445547b_2' }"
 
     input:
     tuple val(meta), path(bam)
 
     output:
-    tuple val(meta), path("*.ccurve.txt"), emit: ccurve
-    tuple val(meta), path("*.log")       , emit: log
-    path  "versions.yml"                 , emit: versions
+    tuple val(meta), path("*.lc_extrap.txt"), emit: lc_extrap
+    tuple val(meta), path("*.log")          , emit: log
+    path  "versions.yml"                    , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
@@ -28,7 +28,7 @@ process PRESEQ_LCEXTRAP {
         lc_extrap \\
         $args \\
         $paired_end \\
-        -output ${prefix}.ccurve.txt \\
+        -output ${prefix}.lc_extrap.txt \\
         $bam
     cp .command.err ${prefix}.command.log
 

modules/preseq/lcextrap/meta.yml

@@ -8,8 +8,8 @@ tools:
   - preseq:
       description: Software for predicting library complexity and genome coverage in high-throughput sequencing
       homepage: http://smithlabresearch.org/software/preseq/
-      documentation: None
-      tool_dev_url: None
+      documentation: http://smithlabresearch.org/wp-content/uploads/manual.pdf
+      tool_dev_url: https://github.com/smithlabcode/preseq
       doi: ""
       licence: ["GPL"]
 
@@ -34,10 +34,10 @@ output:
       type: file
       description: File containing software versions
       pattern: "versions.yml"
-  - ccurve:
+  - lc_extrap:
       type: file
       description: File containing output of Preseq lcextrap
-      pattern: "*.{ccurve.txt}"
+      pattern: "*.{lc_extrap.txt}"
   - log:
       type: file
       description: Log file containing stderr produced by Preseq
@@ -45,3 +45,4 @@ output:
 
 authors:
   - "@drpatelh"
+  - "@Emiller88"

tests/config/pytest_modules.yml

@@ -1367,6 +1367,10 @@ porechop:
   - modules/porechop/**
   - tests/modules/porechop/**
 
+preseq/ccurve:
+  - modules/preseq/ccurve/**
+  - tests/modules/preseq/ccurve/**
+
 preseq/lcextrap:
   - modules/preseq/lcextrap/**
   - tests/modules/preseq/lcextrap/**

tests/modules/preseq/ccurve/main.nf

@@ -0,0 +1,25 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { PRESEQ_CCURVE } from '../../../../modules/preseq/ccurve/main.nf'
+
+workflow test_preseq_ccurve_single_end {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true) 
+    ]
+
+    PRESEQ_CCURVE ( input )
+}
+
+workflow test_preseq_ccurve_paired_end {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true)
+    ]
+
+    PRESEQ_CCURVE ( input )
+}

tests/modules/preseq/ccurve/nextflow.config

@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+}

tests/modules/preseq/ccurve/test.yml

@@ -0,0 +1,19 @@
+- name: preseq ccurve single-end
+  command: nextflow run ./tests/modules/preseq/ccurve -entry test_preseq_ccurve_single_end -c ./tests/config/nextflow.config -c ./tests/modules/preseq/ccurve/nextflow.config
+  tags:
+    - preseq
+    - preseq/ccurve
+  files:
+    - path: output/preseq/test.c_curve.txt
+      md5sum: cf4743abdd355595d6ec1fb3f38e66e5
+    - path: output/preseq/test.command.log
+
+- name: preseq ccurve paired-end
+  command: nextflow run ./tests/modules/preseq/ccurve -entry test_preseq_ccurve_paired_end -c ./tests/config/nextflow.config -c ./tests/modules/preseq/ccurve/nextflow.config
+  tags:
+    - preseq
+    - preseq/ccurve
+  files:
+    - path: output/preseq/test.c_curve.txt
+      md5sum: cf4743abdd355595d6ec1fb3f38e66e5
+    - path: output/preseq/test.command.log

tests/modules/preseq/lcextrap/test.yml

@@ -4,7 +4,7 @@
     - preseq
     - preseq/lcextrap
   files:
-    - path: output/preseq/test.ccurve.txt
+    - path: output/preseq/test.lc_extrap.txt
       md5sum: 1fa5cdd601079329618f61660bee00de
     - path: output/preseq/test.command.log
 
@@ -14,6 +14,6 @@
     - preseq
     - preseq/lcextrap
   files:
-    - path: output/preseq/test.ccurve.txt
+    - path: output/preseq/test.lc_extrap.txt
       md5sum: 10e5ea860e87fb6f5dc10f4f20c62040
     - path: output/preseq/test.command.log