Merge branch 'kraken' into 'master'

Added kraken run + multiqc

See merge request !7

main.nf

@@ -30,6 +30,8 @@ ref_name = file("${ref_path}/${ref_file}").baseName
 primer = file(primer_fasta)
 ref = file("${ref_path}/${ref_file}")
 sum_of_2_nb = fivep_seq_nb.toInteger()+added_nb.toInteger()
+k2db = file("${kraken_db}")
+
 // end from the nextflow.config file
 
 // from parameters
@@ -44,6 +46,7 @@ modules = params.modules // remove the dot -> can be used in bash scripts
 fastq_ch_test = file("${in_path}/${fastq_file}") // to test if exist below
 primer_ch_test = file("${primer_fasta}") // to test if exist below
 ref_ch_test = file("${ref_path}/${ref_file}") // to test if exist below
+k2db_ch_test = file("${kraken_db}") // to test if exist below
 
 //////// end Variables from config.file that need to be checked
 
@@ -70,6 +73,10 @@ if(system_exec == 'local' || system_exec == 'slurm'){
     if( ! file_exists3){
         error "\n\n========\n\nERROR IN NEXTFLOW EXECUTION\n\nINVALID ref_path: ${ref_path}\nOR ref_file: ${ref_file}\nPARAMETER IN nextflow.config FILE\n\nIF POINTING TO A DISTANT SERVER, CHECK THAT IT IS MOUNTED\n\n========\n\n"
     }
+    def file_exists4 = k2db_ch_test.exists()
+    if( ! file_exists4){
+        error "\n\n========\n\nERROR IN NEXTFLOW EXECUTION\n\nINVALID kraken db path: ${kraken_db}\nPARAMETER IN nextflow.config FILE\n\nIF POINTING TO A DISTANT SERVER, CHECK THAT IT IS MOUNTED\n\n========\n\n"
+    }
 }else{
     error "\n\n========\n\nERROR IN NEXTFLOW EXECUTION\n\nINVALID system_exec PARAMETER IN nextflow.config FILE: ${system_exec}\n\n========\n\n"
 }
@@ -140,7 +147,7 @@ process trim { // Trim the oligo sequences. See section 8.4 of the labbook 20200
     val alien_l_param from alientrimmer_l_param
 
     output:
-    file "${file_name}_trim.fq" into fastq_trim_ch1, fastq_trim_ch2
+    file "${file_name}_trim.fq" into fastq_trim_ch1, fastq_trim_ch2, fastq_trim_ch3
     file "report.rmd" into log_ch2
 
     script:
@@ -149,6 +156,22 @@ process trim { // Trim the oligo sequences. See section 8.4 of the labbook 20200
     """
 }
 
+process kraken {
+    label 'kraken'
+
+    input:
+    file fastq from fastq_trim_ch3
+    file k2db
+
+    output:
+    file "${fastq.baseName}_kraken_std.txt" into krakenreports
+
+    script:
+    """
+    kraken2 --db ${k2db} --threads ${task.cpus} --report ${fastq.baseName}_kraken_std.txt ${fastq} > ${fastq.baseName}.kraken2
+    """
+}
+
 
 process fastqc1 { // section 8.5 of the labbook 20200520
     label 'fastqc' // see the withLabel: bash in the nextflow config file 
@@ -419,7 +442,7 @@ process multiQC{
     publishDir "${out_path}/reports", mode: 'copy', overwrite: false //
 
     input:
-    file "*" from fastqc_log_ch.mix(bowtie2_log_ch).collect()
+    file "*" from fastqc_log_ch.mix(bowtie2_log_ch).mix(krakenreports).collect()
 
     output:
     file "*.html"

nextflow.config

@@ -26,6 +26,8 @@ env {
     //fastq_file="test.fastq.gz" // fastq file name
     // fastq_file="Pool-B2699_S1_L001_R1_001.fastq.gz"
     fastq_file="test.fastq2.gz"
+    //kraken_db="$baseDir/assets/k2_standard"
+    kraken_db="/pasteur/zeus/services/p01/banques-prod/prod/rel/kraken_standard/current/kraken/2.1.1/standard/"
     //fastq_file="4-4_S1_L001_R1_001.fastq.gz"
     //primer_fasta="$baseDir/dataset/20200520_adapters_TruSeq_B2699_14985_CL.fasta"
     primer_fasta="$baseDir/dataset/20200520_adapters_TruSeq_B2699_14985_CL.fasta" // list of primers used for the library and used by Alien trimmer to trim the raw reads
@@ -71,7 +73,7 @@ env {
 //////// variables that will be used below (and potentially in the main.nf file)
 
 //// must be also exported
-system_exec = 'local' // the system that runs the workflow. Either 'local' or 'slurm'
+system_exec = 'slurm' // the system that runs the workflow. Either 'local' or 'slurm'
 //docker_exe = true // true for docker and false for singularity
 //out_path="/mnt/c/Users/Gael/Desktop" // where the report file will be saved. Example report_path = '.' for where the main.nf run is executed or report_path = '/mnt/c/Users/Gael/Desktop' 
 out_path="$baseDir/results" // where the report file will be saved. Example report_path = '.' for where the main.nf run is executed or report_path = '/mnt/c/Users/Gael/Desktop' 
@@ -194,6 +196,12 @@ process {
         memory='3G' // only used when name = "local" in the executor part above
     }
 
+    withLabel: kraken{
+        container="evolbioinfo/kraken:v2.0.8-beta"
+        cpus=20
+        memory='100G'
+    }
+
     withLabel: fastqc {
         container='evolbioinfo/fastqc:v0.11.8'
         cpus=1 // only used when name = "local" in the executor part above