Merge pull request #7 from Zymo-Research/zliu-issue6

add downsampling

main.nf

@@ -45,6 +45,9 @@ Channel.from( summary.collect{ [it.key, it.value] } )
  * PROCESS DEFINITION
  */
 include { check_design } from "./processes/check_design"
+include { downsample } from "./processes/downsample" addParams(
+    downsample_num: params.downsample_num
+)
 include { miqscoreShotgun } from "./processes/miqscoreShotgun" addParams(
     publish_dir: "${outdir}/miqscoreShotgun",
 )
@@ -64,7 +67,9 @@ workflow {
             parse_design(it)
         }
         .set { input }
-    miqscoreShotgun(input)
+    downsample(input)
+    miqscore_input = params.downsample_num ? downsample.out.reads : input
+    miqscoreShotgun(miqscore_input)
     miqscoreShotgun.out.report.map { "${outdir}/miqscoreShotgun/" + it.getName() }
         .collectFile(name: "${outdir}/download_data/file_locations.txt", newLine: true)
         .set { output_locations }

nextflow.config

@@ -10,7 +10,8 @@ params {
     // Workflow input
     design = false
 
-    // Input Options
+    // Downsample
+    downsample_num = 2000000
 
     // Max resources, expected to be overwritten if run on your own system
     max_memory = 60.GB

nextflow_schema.json

@@ -57,6 +57,12 @@
             "type": "object",
             "description": "Less common options for the pipeline, typically set in a config file",
             "properties": {
+                "downsample_num": {
+                    "type": "integer",
+                    "description": "Number of read pairs to downsample to. If input has fewer reads than this, downsampling will not happen. Set to 0 to turn this off.",
+                    "minimum": 0,
+                    "default": 2000000
+                },
                 "name": {
                     "type": "string",
                     "description": "Name for the pipeline run. If not specified, Nextflow will automatically generate a random mnemonic",

processes/downsample.nf

@@ -0,0 +1,42 @@
+// Downsample FASTQ files
+
+process downsample {
+    container 'quay.io/biocontainers/seqtk:1.4--he4a0461_1'
+
+    input:
+    tuple val(meta), path(reads)
+
+    when:
+    params.downsample_num
+
+    output:
+    tuple val(meta), path("${meta.name}_downsample_*.fastq.gz"), emit: reads
+
+    script:
+    if (meta.single_end) {
+        """
+        readnum=\$((\$(zcat ${reads[0]} | wc -l) / 4))
+        if ((\$readnum > $params.downsample_num))
+        then
+        seqtk sample -s1000 ${reads[0]} $params.downsample_num > ${meta.name}_downsample_R1.fastq
+        gzip ${meta.name}_downsample_R1.fastq
+        else
+        [ ! -f ${meta.name}_downsample_R1.fastq.gz ] && ln -s ${reads[0]} ${meta.name}_downsample_R1.fastq.gz
+        fi
+        """
+    } else {
+        """
+        readnum=\$((\$(zcat ${reads[0]} | wc -l) / 4))
+        if ((\$readnum > $params.downsample_num))
+        then
+        seqtk sample -s1000 ${reads[0]} $params.downsample_num > ${meta.name}_downsample_R1.fastq
+        gzip ${meta.name}_downsample_R1.fastq
+        seqtk sample -s1000 ${reads[1]} $params.downsample_num > ${meta.name}_downsample_R2.fastq
+        gzip ${meta.name}_downsample_R2.fastq
+        else
+        [ ! -f ${meta.name}_downsample_R1.fastq.gz ] && ln -s ${reads[0]} ${meta.name}_downsample_R1.fastq.gz
+        [ ! -f ${meta.name}_downsample_R2.fastq.gz ] && ln -s ${reads[1]} ${meta.name}_downsample_R2.fastq.gz
+        fi
+        """
+    }
+}