add nextflow pipeline for fastkmers

data/sample.csv

+specie; sex; r1; r2
+mbelari; female; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/1_S1_L001_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/1_S1_L001_R2_001.fasta.gz
+mbelari; female; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/2_S2_L001_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/2_S2_L001_R2_001.fasta.gz
+mbelari; female; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/3_S3_L001_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/3_S3_L001_R2_001.fasta.gz
+mbelari; female; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/4_S4_L001_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/4_S4_L001_R2_001.fasta.gz
+mbelari; female; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/5_S5_L001_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/5_S5_L001_R2_001.fasta.gz
+mbelari; female; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/6_S6_L001_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/6_S6_L001_R2_001.fasta.gz
+mbelari; female; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/MRDR5_R1.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_females/MRDR5_R2.fasta.gz
+mbelari; male; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_males/JU2817_males_S11_L002_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_males/JU2817_males_S11_L002_R2_001.fasta.gz
+mbelari; male; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_males/MRDR6_R1.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mbelari_males/MRDR6_R2.fasta.gz
+mlongespiculosa; female; /scratch/Bio/lmodolo/kmer_diff/data/Mlongespiculosa_females/MRDR3_R1.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mlongespiculosa_females/MRDR3_R2.fasta.gz
+mlongespiculosa; male; /scratch/Bio/lmodolo/kmer_diff/data/Mlongespiculosa_males/MRDR4_R1.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mlongespiculosa_males/MRDR4_R1.fasta.gz
+mspiculigera; female; /scratch/Bio/lmodolo/kmer_diff/data/Mspiculigera_females/AF72-females_BIS_S8_L002_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mspiculigera_females/AF72-females_BIS_S8_L002_R2_001.fasta.gz
+mspiculigera; male; /scratch/Bio/lmodolo/kmer_diff/data/Mspiculigera_males/AF72-males_S7_L002_R1_001.fasta.gz ; /scratch/Bio/lmodolo/kmer_diff/data/Mspiculigera_males/AF72-males_S7_L002_R2_001.fasta.gz

src/main.nf

+nextflow.enable.dsl=2
+
+/*
+Testing pipeline for marseq scRNASeq analysis
+*/
+
+include { FASTKMERS } from "./modules/fastkmers.nf"
+
+
+Channel.fromPath( file(params.csv) )
+                    .splitCsv(header: true)
+                    .map{row ->
+                        fastq_r1 = file("${r1}")
+                        fastq_r2 = file("${r2}")
+                        meta = ['id': ${fastq_r1.simpleName},, 'specie': ${specie}, 'sex': ${sex}]
+                        [meta, [fastq_r1, fastq_r2]]
+                    }
+                    .set{ fastq }
+
+workflow {
+  FASTLMERS(fastq)
+}

src/modules/fastkmers.nf

+/*
+ * normalize bam file
+ */
+process FASTKMERS {
+    tag "$meta.id"
+    label 'process_single'
+
+    container "lbmc/fastkmers:025efdf"
+
+    input:
+    tuple val(meta), path(fastq)
+
+    output:
+    tuple val(meta), path("*.csv"), emit: csv
+    path "versions.yml"           , emit: versions
+
+    script:
+    def args = task.ext.args ?: ''
+    """
+    fastkmers -k 31 ${fastq} > ${fastq.simpleName}.csv
+
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        fastkmers: \$(echo \$(fastkmers --version 2>&1))
+    END_VERSIONS
+    """
+}

src/nextflow.config

+psmn {
+  charliecloud.enabled = true
+  charliecloud.cacheDir = "/Xnfs/abc/charliecloud"
+  charliecloud.readOnlyInputs = true
+  charliecloud.runOptions = "--bind /scratch:/scratch --bind /Xnfs:/Xnfs --bind /home:/home"
+
+  process{
+    errorStrategy = { sleep(Math.pow(2, task.attempt) * 200 as long); return 'retry' }
+    maxRetries = 3
+    executor = "slurm"
+    queue = "Lake"
+    withLabel: big_mem_mono_cpus {
+      cpus = 1
+      memory = "128GB"
+      time = "24h"
+    }
+    withLabel: big_mem_multi_cpus {
+      cpus = 32
+      memory = "192GB"
+      time = "24h"
+    }
+    withLabel: small_mem_mono_cpus {
+      cpus = 1
+      memory = "16GB"
+      time = "24h"
+    }
+    withLabel: small_mem_multi_cpus {
+      cpus = 32
+      memory = "16GB"
+      time = "24h"
+    }
+  }
+
+  params {
+    max_memory = 512.GB
+    max_cpus = 32
+    max_time = 24.h
+  }
+}