Added kallistobustools/ref (#408)

* Added kallistobustools/ref. Local tests all passing with Docker. Linting passed. Test data currently in /tests/data/delete_me

* Removed trailing whitespace line 29

* Moved workflow from meta to options.

* Update main.nf

* Forgot to remove previous testing input channel for workflow.

* Apply suggestions from code review

Applied changes suggested by @drpatelh

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Added gtf to meta.yml.

* Apply suggestions from code review

Adding @drpatelh suggested changes.

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Moved workflow to input value. Fixed tests.

* Update tests/software/kallistobustools/ref/test.yml

* Apply suggestions from code review

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

software/kallistobustools/ref/functions.nf

@@ -0,0 +1,60 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.args3         = args.args3 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/kallistobustools/ref/main.nf

@@ -0,0 +1,67 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process KALLISTOBUSTOOLS_REF {
+    tag "$fasta"
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:'') }
+
+    conda (params.enable_conda ? "bioconda::kb-python=0.25.1" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/kb-python:0.25.1--py_0"
+    } else {
+        container "quay.io/biocontainers/kb-python:0.25.1--py_0"
+    }
+
+    input:
+    path fasta
+    path gtf
+    val  workflow
+
+    output:
+    path "*.version.txt"  , emit: version
+    path "kb_ref_out.idx" , emit: index
+    path "t2g.txt"        , emit: t2g
+    path "cdna.fa"        , emit: cdna
+    path "intron.fa"      , optional:true, emit: intron
+    path "cdna_t2c.txt"   , optional:true, emit: cdna_t2c
+    path "intron_t2c.txt" , optional:true, emit: intron_t2c
+
+    script:
+    def software = getSoftwareName(task.process)
+    if (workflow == "standard") {
+        """
+        kb \\
+            ref \\
+            -i kb_ref_out.idx \\
+            -g t2g.txt \\
+            -f1 cdna.fa \\
+            --workflow $workflow \\
+            $fasta \\
+            $gtf
+
+        echo \$(kb 2>&1) | sed 's/^kb_python //; s/Usage.*\$//' > ${software}.version.txt
+        """
+    } else {
+        """
+        kb \\
+            ref \\
+            -i kb_ref_out.idx \\
+            -g t2g.txt \\
+            -f1 cdna.fa \\
+            -f2 intron.fa \\
+            -c1 cdna_t2c.txt \\
+            -c2 intron_t2c.txt \\
+            --workflow $workflow \\
+            $fasta \\
+            $gtf
+
+        echo \$(kb 2>&1) | sed 's/^kb_python //; s/Usage.*\$//' > ${software}.version.txt
+        """
+    }
+}

software/kallistobustools/ref/meta.yml

@@ -0,0 +1,60 @@
+name: kallistobustools_ref
+description: index creation for kb count quantification of single-cell data.
+keywords:
+  - kallisto-bustools
+  - index
+tools:
+  - kb:
+      description: kallisto|bustools (kb) is a tool developed for fast and efficient processing of single-cell OMICS data.
+      homepage: https://www.kallistobus.tools/
+      documentation: https://kb-python.readthedocs.io/en/latest/index.html
+      tool_dev_url: https://github.com/pachterlab/kb_python
+      doi: "https://doi.org/10.22002/D1.1876"
+      licence: MIT License
+
+input:
+  - fasta:
+      type: file
+      description: Genomic DNA fasta file
+      pattern: "*.{fasta,fasta.gz}"
+  - gtf:
+      type: file
+      description: Genomic gtf file
+      pattern: "*.{gtf,gtf.gz}"
+  - workflow:
+      type: value
+      description: String value defining worfklow to use, can be one of "standard", "lamanno", "nucleus"
+      pattern: "{standard,lamanno,nucleus}"
+
+output:
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - kb_ref_idx:
+      type: file
+      description: Index file from kb ref.
+      pattern: "*.{idx}"
+  - t2g:
+      type: file
+      description: Transcript to gene table
+      pattern: "*t2g.{txt}"
+  - cdna:
+      type: file
+      description: Cdna fasta file
+      pattern: "*cdna.{fa}"
+  - intron:
+      type: file
+      description: intron fasta file
+      pattern: "*intron.{fa}"
+  - cdna_t2c:
+      type: file
+      description: cdna transcript to capture file
+      pattern: "*cdna_t2c.{txt}"
+  - intron_t2c:
+      type: file
+      description: intron transcript to capture file
+      pattern: "*intron_t2c.{txt}"
+
+authors:
+  - "@flowuenne"

tests/config/pytest_software.yml

@@ -182,11 +182,11 @@ fasttree:
   - software/fasttree/**
   - tests/software/fasttree/**
 
-fgbio_callmolecularconsensusreads:
+fgbio/callmolecularconsensusreads:
   - software/fgbio/callmolecularconsensusreads/**
   - tests/software/fgbio/callmolecularconsensusreads/**
 
-fgbio_sortbam:
+fgbio/sortbam:
   - software/fgbio/sortbam/**
   - tests/software/fgbio/sortbam/**
 
@@ -293,6 +293,10 @@ kallisto/index:
   - software/kallisto/index/**
   - tests/software/kallisto/index/**
 
+kallistobustools/ref:
+  - software/kallistobustools/ref/**
+  - tests/software/kallistobustools/ref/**
+
 kraken2/run:
   - software/kraken2/run/**
   - tests/software/kraken2/run/**
@@ -317,11 +321,11 @@ mosdepth:
   - software/mosdepth/**
   - tests/software/mosdepth/**
 
-msisensor_msi:
+msisensor/msi:
   - software/msisensor/msi/**
   - tests/software/msisensor/msi/**
 
-msisensor_scan:
+msisensor/scan:
   - software/msisensor/scan/**
   - tests/software/msisensor/scan/**
 

tests/software/kallistobustools/ref/main.nf

@@ -0,0 +1,32 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { KALLISTOBUSTOOLS_REF } from '../../../../software/kallistobustools/ref/main.nf' addParams( options: [:] )
+
+workflow test_kallistobustools_ref_standard {
+
+    fasta       = file("${launchDir}/tests/data/delete_me/kallistobustools/GRCm39.chr19_100k.fa.gz", checkIfExists: true)
+    gtf         = file("${launchDir}/tests/data/delete_me/kallistobustools/gencode.VM26.chr19_10k.gtf.gz", checkIfExists: true)
+    workflow    = "standard"
+    
+    KALLISTOBUSTOOLS_REF(fasta, gtf, workflow)
+}
+
+workflow test_kallistobustools_ref_lamanno {
+
+    fasta       = file("${launchDir}/tests/data/delete_me/kallistobustools/GRCm39.chr19_100k.fa.gz", checkIfExists: true)
+    gtf         = file("${launchDir}/tests/data/delete_me/kallistobustools/gencode.VM26.chr19_10k.gtf.gz", checkIfExists: true)
+    workflow    = "standard"
+    
+    KALLISTOBUSTOOLS_REF( fasta, gtf, workflow)
+}
+
+workflow test_kallistobustools_ref_nucleus {
+
+    fasta       = file("${launchDir}/tests/data/delete_me/kallistobustools/GRCm39.chr19_100k.fa.gz", checkIfExists: true)
+    gtf         = file("${launchDir}/tests/data/delete_me/kallistobustools/gencode.VM26.chr19_10k.gtf.gz", checkIfExists: true)
+    workflow    = "standard"
+    
+    KALLISTOBUSTOOLS_REF( fasta, gtf, workflow)
+}

tests/software/kallistobustools/ref/test.yml

@@ -0,0 +1,41 @@
+- name: kallistobustools ref test_kallistobustools_ref_standard
+  command: nextflow run tests/software/kallistobustools/ref -entry test_kallistobustools_ref_standard -c tests/config/nextflow.config
+  tags:
+    - kallistobustools
+    - kallistobustools_ref_standard
+    - kallistobustools/ref
+  files:
+    - path: output/kallistobustools/cdna.fa
+      md5sum: da194221c883656c68e2f90d8f77a56b
+    - path: output/kallistobustools/kb_ref_out.idx
+      md5sum: c0ffe9554cb6fdcc7cbb4a1285ca8f29
+    - path: output/kallistobustools/t2g.txt
+      md5sum: 3089d303d9cd1a39a27969eca2dcaba7
+
+- name: kallistobustools ref test_kallistobustools_ref_lamanno
+  command: nextflow run tests/software/kallistobustools/ref -entry test_kallistobustools_ref_lamanno -c tests/config/nextflow.config
+  tags:
+    - kallistobustools
+    - kallistobustools_ref_lamanno
+    - kallistobustools/ref
+  files:
+    - path: output/kallistobustools/cdna.fa
+      md5sum: da194221c883656c68e2f90d8f77a56b
+    - path: output/kallistobustools/kb_ref_out.idx
+      md5sum: c0ffe9554cb6fdcc7cbb4a1285ca8f29
+    - path: output/kallistobustools/t2g.txt
+      md5sum: 3089d303d9cd1a39a27969eca2dcaba7
+
+- name: kallistobustools ref test_kallistobustools_ref_nucleus
+  command: nextflow run tests/software/kallistobustools/ref -entry test_kallistobustools_ref_nucleus -c tests/config/nextflow.config
+  tags:
+    - kallistobustools_ref_nucleus
+    - kallistobustools
+    - kallistobustools/ref
+  files:
+    - path: output/kallistobustools/cdna.fa
+      md5sum: da194221c883656c68e2f90d8f77a56b
+    - path: output/kallistobustools/kb_ref_out.idx
+      md5sum: c0ffe9554cb6fdcc7cbb4a1285ca8f29
+    - path: output/kallistobustools/t2g.txt
+      md5sum: 3089d303d9cd1a39a27969eca2dcaba7