add module for sistr (#1323)

* add module for sistr

* Update test.yml

modules/sistr/main.nf

@@ -0,0 +1,49 @@
+process SISTR {
+    tag "$meta.id"
+    label 'process_medium'
+
+    conda (params.enable_conda ? "bioconda::sistr_cmd=1.1.1" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/sistr_cmd:1.1.1--pyh864c0ab_2':
+        'quay.io/biocontainers/sistr_cmd:1.1.1--pyh864c0ab_2' }"
+
+    input:
+    tuple val(meta), path(fasta)
+
+    output:
+    tuple val(meta), path("*.tab")         , emit: tsv
+    tuple val(meta), path("*-allele.fasta"), emit: allele_fasta
+    tuple val(meta), path("*-allele.json") , emit: allele_json
+    tuple val(meta), path("*-cgmlst.csv")  , emit: cgmlst_csv
+    path "versions.yml"                    , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def is_compressed = fasta.getName().endsWith(".gz") ? true : false
+    def fasta_name = fasta.getName().replace(".gz", "")
+    """
+    if [ "$is_compressed" == "true" ]; then
+        gzip -c -d $fasta > $fasta_name
+    fi
+
+    sistr \\
+        --qc \\
+        $args \\
+        --threads $task.cpus \\
+        --alleles-output ${prefix}-allele.json \\
+        --novel-alleles ${prefix}-allele.fasta \\
+        --cgmlst-profiles ${prefix}-cgmlst.csv \\
+        --output-prediction ${prefix} \\
+        --output-format tab \\
+        $fasta_name
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        sistr: \$(echo \$(sistr --version 2>&1) | sed 's/^.*sistr_cmd //; s/ .*\$//' )
+    END_VERSIONS
+    """
+}

modules/sistr/meta.yml

@@ -0,0 +1,55 @@
+name: sistr
+description: Serovar prediction of salmonella assemblies
+keywords:
+  - bacteria
+  - fasta
+  - salmonella
+tools:
+  - sistr:
+      description: Salmonella In Silico Typing Resource (SISTR) commandline tool for serovar prediction
+      homepage: https://github.com/phac-nml/sistr_cmd
+      documentation: https://github.com/phac-nml/sistr_cmd
+      tool_dev_url: https://github.com/phac-nml/sistr_cmd
+      doi: "10.1371/journal.pone.0147101"
+      licence: ['Apache-2.0']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - fasta:
+      type: file
+      description: Nucleotide or protein sequences in FASTA format
+      pattern: "*.{fasta,fasta.gz,fa,fa.gz,fna,fna.gz,faa,faa.gz}"
+
+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"
+  - tsv:
+      type: file
+      description: SISTR serovar prediction
+      pattern: "*.{tsv}"
+  - allele_json:
+      type: file
+      description: Allele sequences and info to JSON
+      pattern: "*.{json}"
+  - allele_fasta:
+      type: file
+      description: FASTA file destination of novel cgMLST alleles
+      pattern: "*.{fasta}"
+  - cgmlst_csv:
+      type: file
+      description: CSV file destination for cgMLST allelic profiles
+      pattern: "*.{csv}"
+
+authors:
+  - "@rpetit3"

tests/config/pytest_modules.yml

@@ -1429,6 +1429,10 @@ shovill:
   - modules/shovill/**
   - tests/modules/shovill/**
 
+sistr:
+  - modules/sistr/**
+  - tests/modules/sistr/**
+
 snpdists:
   - modules/snpdists/**
   - tests/modules/snpdists/**

tests/modules/sistr/main.nf

@@ -0,0 +1,15 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SISTR } from '../../../modules/sistr/main.nf'
+
+workflow test_sistr {
+    
+    input = [ 
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['haemophilus_influenzae']['genome']['genome_fna_gz'], checkIfExists: true) 
+    ]
+
+    SISTR ( input )
+}

tests/modules/sistr/nextflow.config

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

tests/modules/sistr/test.yml

@@ -0,0 +1,15 @@
+- name: sistr test_sistr
+  command: nextflow run tests/modules/sistr -entry test_sistr -c tests/config/nextflow.config
+  tags:
+    - sistr
+  files:
+    - path: output/sistr/test-allele.fasta
+      md5sum: 144a74999eb9dd01520be5c61e8bd210
+    - path: output/sistr/test-allele.json
+      md5sum: 3eb993c9489904621f539a93ff9a90ec
+    - path: output/sistr/test-cgmlst.csv
+      md5sum: c50a2144955fe1b98a6d5792bf295088
+    - path: output/sistr/test.tab
+      contains: ["cgmlst_ST", "serovar", "matched"]
+    - path: output/sistr/versions.yml
+      md5sum: 8b852f002c3ce67e3f6498da15b28296