rna-seq.nf

#!/usr/bin/env nextflow

//  ########  ####  ######  ########
//  ##     ##  ##  ##    ## ##
//  ##     ##  ##  ##       ##
//  ########   ##  ##       ######
//  ##     ##  ##  ##       ##
//  ##     ##  ##  ##    ## ##
//  ########  ####  ######  ##

// Define input variables
params.deriva = "${baseDir}/../test_data/auth/credential.json"
params.bdbag = "${baseDir}/../test_data/auth/cookies.txt"
//params.repRID = "16-1ZX4"
params.repRID = "Q-Y5F6"
params.source = "dev"
params.refMoVersion = "38.p6.vM25"
params.refHuVersion = "38.p13.v36"
params.refERCCVersion = "92"
params.outDir = "${baseDir}/../output"
params.upload = false
params.email = ""


// Define override input variable
params.refSource = "biohpc"
params.inputBagForce = ""
params.fastqsForce = ""
params.speciesForce = ""

// Define tracking input variables
params.ci = false
params.dev = true


// Parse input variables
deriva = Channel
  .fromPath(params.deriva)
  .ifEmpty { exit 1, "deriva credential file not found: ${params.deriva}" }
deriva.into {
  deriva_getBag
  deriva_getRefInfer
  deriva_getRef
  deriva_uploadInputBag
  deriva_uploadExecutionRun
  deriva_uploadQC
  deriva_uploadProcessedFile
  deriva_uploadOutputBag
  deriva_finalizeExecutionRun
  deriva_failPreExecutionRun
  deriva_failExecutionRun
}
bdbag = Channel
  .fromPath(params.bdbag)
  .ifEmpty { exit 1, "deriva cookie file for bdbag not found: ${params.bdbag}" }
repRID = params.repRID
refMoVersion = params.refMoVersion
refHuVersion = params.refHuVersion
refERCCVersion = params.refERCCVersion
outDir = params.outDir
logsDir = "${outDir}/Logs"
upload = params.upload
inputBagForce = params.inputBagForce
fastqsForce = params.fastqsForce
speciesForce = params.speciesForce
email = params.email

// Define fixed files and variables
replicateExportConfig = Channel.fromPath("${baseDir}/conf/Replicate_For_Input_Bag.json")
executionRunExportConfig = Channel.fromPath("${baseDir}/conf/Execution_Run_For_Output_Bag.json")
if (params.source == "dev") {
  source = "dev.gudmap.org"
} else if (params.source == "staging") {
  source = "staging.gudmap.org"
} else if (params.source == "production") {
  source = "www.gudmap.org"
}
if (params.refSource == "biohpc") {
  referenceBase = "/project/BICF/BICF_Core/shared/gudmap/references/new"
} else if (params.refSource == "datahub") {
  referenceBase = "www.gudmap.org"
}
referenceInfer = Channel.fromList(["ERCC","GRCh","GRCm"])
multiqcConfig = Channel.fromPath("${baseDir}/conf/multiqc_config.yaml")
bicfLogo = Channel.fromPath("${baseDir}/../docs/bicf_logo.png")
softwareReferences = Channel.fromPath("${baseDir}/../docs/software_references_mqc.yaml")
softwareVersions = Channel.fromPath("${baseDir}/../docs/software_versions_mqc.yaml")

// Define script files
script_bdbagFetch = Channel.fromPath("${baseDir}/scripts/bdbag_fetch.sh")
script_parseMeta = Channel.fromPath("${baseDir}/scripts/parse_meta.py")
script_inferMeta = Channel.fromPath("${baseDir}/scripts/infer_meta.sh")
script_refDataInfer = Channel.fromPath("${baseDir}/scripts/extract_ref_data.py")
script_refData = Channel.fromPath("${baseDir}/scripts/extract_ref_data.py")
script_calculateTPM = Channel.fromPath("${baseDir}/scripts/calculateTPM.R")
script_convertGeneSymbols = Channel.fromPath("${baseDir}/scripts/convertGeneSymbols.R")
script_tinHist = Channel.fromPath("${baseDir}/scripts/tin_hist.py")
script_uploadInputBag = Channel.fromPath("${baseDir}/scripts/upload_input_bag.py")
script_uploadExecutionRun_uploadExecutionRun = Channel.fromPath("${baseDir}/scripts/upload_execution_run.py")
script_uploadExecutionRun_finalizeExecutionRun = Channel.fromPath("${baseDir}/scripts/upload_execution_run.py")
script_uploadExecutionRun_failPreExecutionRun = Channel.fromPath("${baseDir}/scripts/upload_execution_run.py")
script_uploadExecutionRun_failExecutionRun = Channel.fromPath("${baseDir}/scripts/upload_execution_run.py")
script_uploadQC = Channel.fromPath("${baseDir}/scripts/upload_qc.py")
script_uploadOutputBag = Channel.fromPath("${baseDir}/scripts/upload_output_bag.py")
script_deleteEntry_uploadQC = Channel.fromPath("${baseDir}/scripts/delete_entry.py")
script_deleteEntry_uploadProcessedFile = Channel.fromPath("${baseDir}/scripts/delete_entry.py")

/*
 * trackStart: track start of pipeline
 */
process trackStart {
  container 'docker://gudmaprbk/gudmap-rbk_base:1.0.0'
  script:
  """
  hostname
  ulimit -a

  curl -H 'Content-Type: application/json' -X PUT -d \
    '{ \
      "sessionId": "${workflow.sessionId}", \
      "pipeline": "gudmap.rbk_rnaseq", \
      "start": "${workflow.start}", \
      "repRID": "${repRID}", \
      "astrocyte": false, \
      "status": "started", \
      "nextflowVersion": "${workflow.nextflow.version}", \
      "pipelineVersion": "${workflow.manifest.version}", \
      "ci": ${params.ci}, \
      "dev": ${params.dev} \
    }' \
    "https://xku43pcwnf.execute-api.us-east-1.amazonaws.com/ProdDeploy/pipeline-tracking"
  """
}

log.info """\
====================================
BICF RNA-seq Pipeline for GUDMAP/RBK
====================================
Replicate RID          : ${params.repRID}
Source Server          : ${params.source}
Mouse Reference Version: ${params.refMoVersion}
Human Reference Version: ${params.refHuVersion}
ERCC Reference Version : ${params.refERCCVersion}
Reference source       : ${params.refSource}
Output Directory       : ${params.outDir}
Upload                 : ${upload}
------------------------------------
Nextflow Version       : ${workflow.nextflow.version}
Pipeline Version       : ${workflow.manifest.version}
Session ID             : ${workflow.sessionId}
------------------------------------
CI                     : ${params.ci}
Development            : ${params.dev}
------------------------------------
"""

/*
 * getBag: download input bag
 */
process getBag {
  tag "${repRID}"
  publishDir "${outDir}/inputBag", mode: 'copy', pattern: "*_inputBag_*.zip"

  input:
    path credential, stageAs: "credential.json" from deriva_getBag
    path replicateExportConfig

  output:
    path ("*.zip") into bag

  when:
    inputBagForce == ""

  script:
    """
    hostname > ${repRID}.getBag.log
    ulimit -a >> ${repRID}.getBag.log

    # link credential file for authentication
    echo -e "LOG: linking deriva credentials" >> ${repRID}.getBag.log
    mkdir -p ~/.deriva
    ln -sf `readlink -e credential.json` ~/.deriva/credential.json
    echo -e "LOG: linked" >> ${repRID}.getBag.log

    # deriva-download replicate RID
    echo -e "LOG: fetching bag for ${repRID} in GUDMAP" >> ${repRID}.getBag.log
    deriva-download-cli ${source} --catalog 2 ${replicateExportConfig} . rid=${repRID}
    echo -e "LOG: fetched" >> ${repRID}.getBag.log

    name=\$(ls *.zip)
    name=\$(basename \${name} | cut -d "." -f1)
    yr=\$(date +'%Y')
    mn=\$(date +'%m')
    dy=\$(date +'%d')
    mv \${name}.zip \${name}_\${yr}\${mn}\${dy}.zip
    """
}

// Set inputBag to downloaded or forced input
if (inputBagForce != "") {
  inputBag = Channel
    .fromPath(inputBagForce)
    .ifEmpty { exit 1, "override inputBag file not found: ${inputBagForce}" }
} else {
  inputBag = bag
}
inputBag.into {
  inputBag_getData
  inputBag_uploadInputBag
}

/*
 * getData: fetch replicate files from consortium with downloaded bdbag.zip
 */
process getData {
  tag "${repRID}"

  input:
    path script_bdbagFetch
    path cookies, stageAs: "deriva-cookies.txt" from bdbag
    path inputBag from inputBag_getData

  output:
    path ("*.R{1,2}.fastq.gz") into fastqs
    path ("**/File.csv") into fileMeta
    path ("**/Experiment Settings.csv") into experimentSettingsMeta
    path ("**/Experiment.csv") into experimentMeta
    path "fastqCount.csv" into fastqCount_fl

  script:
    """
    hostname > ${repRID}.getData.log
    ulimit -a >> ${repRID}.getData.log

    # link deriva cookie for authentication
    echo -e "LOG: linking deriva cookie" >> ${repRID}.getData.log
    mkdir -p ~/.bdbag
    ln -sf `readlink -e deriva-cookies.txt` ~/.bdbag/deriva-cookies.txt
    echo -e "LOG: linked" >> ${repRID}.getData.log

    # get bag basename
    replicate=\$(basename "${inputBag}")
    echo -e "LOG: bag replicate name \${replicate}" >> ${repRID}.getData.log

    # unzip bag
    echo -e "LOG: unzipping replicate bag" >> ${repRID}.getData.log
    unzip ${inputBag}
    echo -e "LOG: unzipped" >> ${repRID}.getData.log

    # bag fetch fastq's only and rename by repRID
    echo -e "LOG: fetching replicate bdbag" >> ${repRID}.getData.log
    sh ${script_bdbagFetch} \${replicate::-13} ${repRID}
    echo -e "LOG: fetched" >> ${repRID}.getData.log
    
    fastqCount=\$(ls *.fastq.gz | wc -l)
    echo "\${fastqCount}" > fastqCount.csv
    """
}

// Split fastq count into channel
fastqCount = Channel.create()
fastqCount_fl.splitCsv(sep: ",", header: false).separate(
  fastqCount
)

// Set raw fastq to downloaded or forced input and replicate them for multiple process inputs
if (fastqsForce != "") {
  Channel
    .fromPath(fastqsForce)
    .ifEmpty { exit 1, "override inputBag file not found: ${fastqsForce}" }
    .collect().into {
      fastqs_parseMetadata
      fastqs_trimData
    }
} else {
  fastqs.into {
    fastqs_parseMetadata
    fastqs_trimData
  }
}

/*
 * parseMetadata: parses metadata to extract experiment parameters
*/
process parseMetadata {
  tag "${repRID}"

  input:
    path script_parseMeta
    path file from fileMeta
    path experimentSettings, stageAs: "ExperimentSettings.csv" from experimentSettingsMeta
    path experiment from experimentMeta
    path (fastq) from fastqs_parseMetadata
    val fastqCount

  output:
    path "design.csv" into metadata_fl
    path "fastqError.csv" into fastqError_fl

  script:
    """
    hostname > ${repRID}.parseMetadata.log
    ulimit -a >> ${repRID}.parseMetadata.log

    # check replicate RID metadata
    rep=\$(python3 ${script_parseMeta} -r ${repRID} -m "${file}" -p repRID)
    echo -e "LOG: replicate RID metadata parsed: \${rep}" >> ${repRID}.parseMetadata.log

    # get experiment RID metadata
    exp=\$(python3 ${script_parseMeta} -r ${repRID} -m "${file}" -p expRID)
    echo -e "LOG: experiment RID metadata parsed: \${exp}" >> ${repRID}.parseMetadata.log

    # get study RID metadata
    study=\$(python3 ${script_parseMeta} -r ${repRID} -m "${file}" -p studyRID)
    echo -e "LOG: study RID metadata parsed: \${study}" >> ${repRID}.parseMetadata.log

    # get endedness metadata
    endsRaw=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experimentSettings}" -p endsMeta)
    echo -e "LOG: endedness metadata parsed: \${endsRaw}" >> ${repRID}.parseMetadata.log
    if [ "\${endsRaw}" == "Single Read" ]
    then
      endsMeta="se"
    elif [ "\${endsRaw}" == "Paired End" ]
    then
      endsMeta="pe"
    else
      endsMeta="unknown"
    fi
    if [ "\${endsRaw}" == "" ]
    then
      endsRaw="_No value_"
    fi

    # ganually get endness
    endsManual=\$(python3 ${script_parseMeta} -r ${repRID} -m "${file}" -p endsManual)
    echo -e "LOG: endedness manually detected: \${endsManual}" >> ${repRID}.parseMetadata.log

    # get strandedness metadata
    stranded=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experimentSettings}" -p stranded)
    echo -e "LOG: strandedness metadata parsed: \${stranded}" >> ${repRID}.parseMetadata.log

    # get spike-in metadata
    spike=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experimentSettings}" -p spike)
    echo -e "LOG: spike-in metadata parsed: \${spike}" >> ${repRID}.parseMetadata.log

    # get species metadata
    species=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experiment}" -p species)
    echo -e "LOG: species metadata parsed: \${species}" >> ${repRID}.parseMetadata.log

    # get read length metadata
    readLength=\$(python3 ${script_parseMeta} -r ${repRID} -m "${experimentSettings}" -p readLength)
    if [ "\${readLength}" = "nan" ]
    then
      readLength="NA"
    fi
    echo -e "LOG: read length metadata parsed: \${readLength}" >> ${repRID}.parseMetadata.log

    # check not incorrect number of fastqs
    fastqCountError=false
    fastqCountError_details=""
    if [ "${fastqCount}" -gt "2" ]
    then
      fastqCountError=true
      fastqCountError_details="**Too many fastqs detected (>2)**"
    elif [ "\${endsMeta}" == "se" ] && [ "${fastqCount}" -ne "1" ]
    then
      fastqCountError=true
      fastqCountError_details="**Number of fastqs detected does not match submitted endness**"
    elif [ "\${endsMeta}" == "pe" ] && [ "${fastqCount}" -ne "2" ]
    then
      fastqCountError=true
      fastqCountError_details="**Number of fastqs detected does not match submitted endness**"
    fi

    # check read counts match for fastqs
    fastqReadError=false
    fastqReadError_details=""
    if [ "\${endsManual}" == "pe" ]
    then
      r1Count=\$(zcat ${fastq[0]} | wc -l)
      r2Count=\$(zcat ${fastq[1]} | wc -l)
      if [ "\${r1Count}" -ne "\${r2Count}" ]
      then
        fastqReadError=true
        fastqReadError_details="**Number of reads do not match for R1 and R2:** there may be a trunkation or mismatch of fastq files"
      fi
    fi

    # save design file
    echo "\${endsMeta},\${endsRaw},\${endsManual},\${stranded},\${spike},\${species},\${readLength},\${exp},\${study}" > design.csv

    # save fastq error file
    echo "\${fastqCountError},\${fastqCountError_details},\${fastqReadError},\${fastqReadError_details}" > fastqError.csv
    """
}

// Split metadata into separate channels
endsMeta = Channel.create()
endsRaw = Channel.create()
endsManual = Channel.create()
strandedMeta = Channel.create()
spikeMeta = Channel.create()
speciesMeta = Channel.create()
readLengthMeta = Channel.create()
expRID = Channel.create()
studyRID = Channel.create()
metadata_fl.splitCsv(sep: ",", header: false).separate(
  endsMeta,
  endsRaw,
  endsManual,
  strandedMeta,
  spikeMeta,
  speciesMeta,
  readLengthMeta,
  expRID,
  studyRID
)

// Replicate metadata for multiple process inputs
endsMeta.into {
  endsMeta_checkMetadata
  endsMeta_aggrQC
  endsMeta_failExecutionRun
}
endsManual.into {
  endsManual_trimData
  endsManual_downsampleData
  endsManual_alignSampleData
  endsManual_aggrQC
}
strandedMeta.into {
  strandedMeta_checkMetadata
  strandedMeta_aggrQC
  strandedMeta_failExecutionRun
}
spikeMeta.into {
  spikeMeta_checkMetadata
  spikeMeta_aggrQC
  spikeMeta_failPreExecutionRun
  spikeMeta_failExecutionRun
}
speciesMeta.into {
  speciesMeta_checkMetadata
  speciesMeta_aggrQC
  speciesMeta_failPreExecutionRun
  speciesMeta_failExecutionRun
}
studyRID.into {
  studyRID_aggrQC
  studyRID_uploadInputBag
  studyRID_uploadProcessedFile
  studyRID_uploadOutputBag
}
expRID.into {
  expRID_aggrQC
  expRID_uploadProcessedFile
}

// Split fastq count error into separate channel
fastqCountError = Channel.create()
fastqCountError_details = Channel.create()
fastqReadError = Channel.create()
fastqReadError_details = Channel.create()
fastqError_fl.splitCsv(sep: ",", header: false).separate(
  fastqCountError,
  fastqCountError_details,
  fastqReadError,
  fastqReadError_details
)

//  Replicate errors for multiple process inputs
fastqCountError.into {
  fastqCountError_trimData
  fastqCountError_getRefInfer
  fastqCountError_downsampleData
  fastqCountError_alignSampleData
  fastqCountError_inferMetadata
  fastqCountError_checkMetadata
  fastqCountError_uploadExecutionRun
  fastqCountError_getRef
  fastqCountError_alignData
  fastqCountError_dedupData
  fastqCountError_makeBigWig
  fastqCountError_countData
  fastqCountError_fastqc
  fastqCountError_dataQC
  fastqCountError_aggrQC
  fastqCountError_uploadQC
  fastqCountError_uploadProcessedFile
  fastqCountError_uploadOutputBag
  fastqCountError_failPreExecutionRun
}
fastqReadError.into {
  fastqReadError_trimData
  fastqReadError_getRefInfer
  fastqReadError_downsampleData
  fastqReadError_alignSampleData
  fastqReadError_inferMetadata
  fastqReadError_checkMetadata
  fastqReadError_uploadExecutionRun
  fastqReadError_getRef
  fastqReadError_alignData
  fastqReadError_dedupData
  fastqReadError_makeBigWig
  fastqReadError_countData
  fastqReadError_fastqc
  fastqReadError_dataQC
  fastqReadError_aggrQC
  fastqReadError_uploadQC
  fastqReadError_uploadProcessedFile
  fastqReadError_uploadOutputBag
  fastqReadError_failPreExecutionRun
}

/*
 * trimData: trims any adapter or non-host sequences from the data
*/
process trimData {
  tag "${repRID}"

  input:
    path (fastq) from fastqs_trimData
    val ends from endsManual_trimData
    val fastqCountError_trimData
    val fastqReadError_trimData

  output:
    path ("*.fq.gz") into fastqsTrim
    path ("*.fastq.gz", includeInputs:true) into fastqs_fastqc
    path ("*_trimming_report.txt") into trimQC
    path ("readLength.csv") into readLengthInfer_fl

  when:
    fastqCountError_trimData == "false"
    fastqReadError_trimData == "false"

  script:
    """
    hostname > ${repRID}.trimData.log
    ulimit -a >> ${repRID}.trimData.log

    # trim fastq's using trim_galore and extract median read length
    echo -e "LOG: trimming ${ends}" >> ${repRID}.trimData.log
    if [ "${ends}" == "se" ]
    then
      trim_galore --gzip -q 25 --length 35 --basename ${repRID} ${fastq[0]}
      readLength=\$(zcat *_trimmed.fq.gz | awk '{if(NR%4==2) print length(\$1)}' | sort -n | awk '{a[NR]=\$0}END{print(NR%2==1)?a[int(NR/2)+1]:(a[NR/2]+a[NR/2+1])/2}')
    elif [ "${ends}" == "pe" ]
    then
      trim_galore --gzip -q 25 --length 35 --paired --basename ${repRID} ${fastq[0]} ${fastq[1]}
      readLength=\$(zcat *_1.fq.gz | awk '{if(NR%4==2) print length(\$1)}' | sort -n | awk '{a[NR]=\$0}END{print(NR%2==1)?a[int(NR/2)+1]:(a[NR/2]+a[NR/2+1])/2}')
    fi
    echo -e "LOG: trimmed" >> ${repRID}.trimData.log
    echo -e "LOG: average trimmed read length: \${readLength}" >> ${repRID}.trimData.log

    # save read length file
    echo "\${readLength}" > readLength.csv
    """
}

// Extract calculated read length metadata into channel
readLengthInfer = Channel.create()
readLengthInfer_fl.splitCsv(sep: ",", header: false).separate(
  readLengthInfer
)

// Replicate inferred read length for multiple process inputs
readLengthInfer.into {
  readLengthInfer_aggrQC
  readLengthInfer_uploadQC
}
// Replicate trimmed fastq's for multiple process inputs
fastqsTrim.into {
  fastqsTrim_alignData
  fastqsTrim_downsampleData
}

// Combine inputs of getRefInfer
getRefInferInput = referenceInfer.combine(deriva_getRefInfer.combine(script_refDataInfer.combine(fastqCountError_getRefInfer.combine(fastqReadError_getRefInfer))))

/*
  * getRefInfer: dowloads appropriate reference for metadata inference
*/
process getRefInfer {
  tag "${refName}"

  input:
    tuple val (refName), path (credential, stageAs: "credential.json"), path (script_refDataInfer), val (fastqCountError), val (fastqReadError) from getRefInferInput

  output:
    tuple val (refName), path ("hisat2", type: 'dir'), path ("*.fna"), path ("*.gtf")  into refInfer
    path ("${refName}", type: 'dir') into bedInfer

  when:
    fastqCountError == "false"
    fastqReadError == "false"

  script:
    """
    hostname > ${repRID}.${refName}.getRefInfer.log
    ulimit -a >> ${repRID}.${refName}.getRefInfer.log

    # link credential file for authentication
    echo -e "LOG: linking deriva credentials" >> ${repRID}.${refName}.getRefInfer.log
    mkdir -p ~/.deriva
    ln -sf `readlink -e credential.json` ~/.deriva/credential.json
    echo -e "LOG: linked" >> ${repRID}.${refName}.getRefInfer.log

    # set the reference name
    if [ "${refName}" == "ERCC" ]
    then
      references=\$(echo ${referenceBase}/ERCC${refERCCVersion})
    elif [ "${refName}" == "GRCm" ]
    then
      references=\$(echo ${referenceBase}/GRCm${refMoVersion})
    elif [ '${refName}' == "GRCh" ]
    then
      references=\$(echo ${referenceBase}/GRCh${refHuVersion})
    else
      echo -e "LOG: ERROR - References could not be set!\nReference found: ${referenceBase}" >> ${repRID}.${refName}.getRefInfer.log
      exit 1
    fi

    # retreive appropriate reference appropriate location
    echo -e "LOG: fetching ${refName} reference files from ${referenceBase}" >> ${repRID}.${refName}.getRefInfer.log
    if [ ${referenceBase} == "/project/BICF/BICF_Core/shared/gudmap/references/new" ]
    then
      unzip \${references}.zip
      mv \$(basename \${references})/data/* .
    elif [ params.refSource == "datahub" ]
    then
      GRCv=\$(echo \${references} | grep -o ${refName}.* | cut -d '.' -f1)
      GRCp=\$(echo \${references} | grep -o ${refName}.* | cut -d '.' -f2)
      GENCODE=\$(echo \${references} | grep -o ${refName}.* | cut -d '.' -f3)
      if [ "${refName}" != "ERCC" ]
      then
        query=\$(echo 'https://${referenceBase}/ermrest/catalog/2/entity/RNASeq:Reference_Genome/Reference_Version='\${GRCv}'.'\${GRCp}'/Annotation_Version=GENCODE%20'\${GENCODE})
      else
        query=\$(echo 'https://${referenceBase}/ermrest/catalog/2/entity/RNASeq:Reference_Genome/Reference_Version=${refName}${refERCCVersion}/Annotation_Version=${refName}${refERCCVersion}')
      fi
      curl --request GET \${query} > refQuery.json
      refURL=\$(python ${script_refDataInfer} --returnParam URL)
      loc=\$(dirname \${refURL})
      fName=\$(python ${script_refDataInfer} --returnParam fName)
      fName=\${fName%.*}
      if [ "\${loc}" = "/hatrac/*" ]; then echo "LOG: Reference not present in hatrac"; exit 1; fi
      filename=\$(echo \$(basename \${refURL}) | grep -oP '.*(?=:)')
      deriva-hatrac-cli --host ${referenceBase} get \${refURL}
      unzip \$(basename \${refURL})
      mv \${fName}/data/* .
    fi
    mv ./annotation/genome.gtf .
    mv ./sequence/genome.fna .
    mkdir ${refName}
    if [ "${refName}" != "ERCC" ]
    then
      mv ./annotation/genome.bed ./${refName}
    fi
    echo -e "LOG: fetched" >> ${repRID}.${refName}.getRefInfer.log
    """
}

/*
 * downsampleData: downsample fastq's for metadata inference
 */
process downsampleData {
  tag "${repRID}"

  input:
    path fastq from fastqsTrim_downsampleData
    val ends from endsManual_downsampleData
    val fastqCountError_downsampleData
    val fastqReadError_downsampleData

  output:
    path ("sampled.1.fq") into fastqs1Sample
    path ("sampled.2.fq") into fastqs2Sample

  when:
    fastqCountError_downsampleData == "false"
    fastqReadError_downsampleData == "false"

  script:
    """
    hostname > ${repRID}.downsampleData.log
    ulimit -a >> ${repRID}.downsampleData.log

    if [ "${ends}" == "se" ]
    then
      echo -e "LOG: downsampling SE trimmed fastq" >> ${repRID}.downsampleData.log
      seqtk sample -s100 *trimmed.fq.gz 100000 1> sampled.1.fq
      touch sampled.2.fq
    elif [ "${ends}" == "pe" ]
    then
      echo -e "LOG: downsampling R1 of PE trimmed fastq" >> ${repRID}.downsampleData.log
      seqtk sample -s100 *1.fq.gz 1000000 1> sampled.1.fq
      echo -e "LOG: downsampling R2 of PE trimmed fastq" >> ${repRID}.downsampleData.log
      seqtk sample -s100 *2.fq.gz 1000000 1> sampled.2.fq
    fi
    echo -e "LOG: downsampled" >> ${repRID}.downsampleData.log
    """
}

// Replicate the dowsampled fastq's and attatched to the references
inferInput = endsManual_alignSampleData.combine(refInfer.combine(fastqs1Sample.collect().combine(fastqs2Sample.collect().combine(fastqCountError_alignSampleData.combine(fastqReadError_alignSampleData)))))

/*
 * alignSampleData: aligns the downsampled reads to a reference database
*/
process alignSampleData {
  tag "${ref}"

  input:
    tuple val (ends), val (ref), path (hisat2), path (fna), path (gtf), path (fastq1), path (fastq2), val (fastqCountError), val (fastqReadError) from inferInput

  output:
    path ("${ref}.sampled.sorted.bam") into sampleBam
    path ("${ref}.sampled.sorted.bam.bai") into sampleBai
    path ("${ref}.alignSampleSummary.txt") into alignSampleQC

  when:
    fastqCountError == "false"
    fastqReadError == "false"

  script:
    """
    hostname > ${repRID}.${ref}.alignSampleData.log
    ulimit -a >> ${repRID}.${ref}.alignSampleData.log

    # align the reads with Hisat2
    echo -e "LOG: aligning ${ends}" >> ${repRID}.${ref}.alignSampleData.log
    if [ "${ends}" == "se" ]
    then

      hisat2 -p `nproc` --add-chrname -S ${ref}.sampled.sam -x hisat2/genome -U ${fastq1} --summary-file ${ref}.alignSampleSummary.txt --new-summary
    elif [ "${ends}" == "pe" ]
    then
      hisat2 -p `nproc` --add-chrname -S ${ref}.sampled.sam -x hisat2/genome --no-mixed --no-discordant -1 ${fastq1} -2 ${fastq2} --summary-file ${ref}.alignSampleSummary.txt --new-summary
    fi
    echo -e "LOG: aliged" >> ${repRID}.${ref}.alignSampleData.log

    # convert the output sam file to a sorted bam file using Samtools
    echo -e "LOG: converting from sam to bam" >> ${repRID}.${ref}.alignSampleData.log
    samtools view -1 -@ `nproc` -F 4 -F 8 -F 256 -o ${ref}.sampled.bam ${ref}.sampled.sam

    # sort the bam file using Samtools
    echo -e "LOG: sorting the bam file" >> ${repRID}.${ref}.alignSampleData.log
    samtools sort -@ `nproc` -O BAM -o ${ref}.sampled.sorted.bam ${ref}.sampled.bam

    # index the sorted bam using Samtools
    echo -e "LOG: indexing sorted bam file" >> ${repRID}.${ref}.alignSampleData.log
    samtools index -@ `nproc` -b ${ref}.sampled.sorted.bam ${ref}.sampled.sorted.bam.bai
    """
}

alignSampleQC.into {
  alignSampleQC_inferMetadata
  alignSampleQC_aggrQC
}

process inferMetadata {
  tag "${repRID}"

  input:
    path script_inferMeta
    path beds from bedInfer.collect()
    path bam from sampleBam.collect()
    path bai from sampleBai.collect()
    path alignSummary from alignSampleQC_inferMetadata.collect()
    val fastqCountError_inferMetadata
    val fastqReadError_inferMetadata

  output:
    path "infer.csv" into inferMetadata_fl
    path "${repRID}.infer_experiment.txt" into inferExperiment
    path "speciesError.csv" into speciesError_fl

  when:
    fastqCountError_inferMetadata == "false"
    fastqReadError_inferMetadata == "false"

  script:
    """
    hostname > ${repRID}.inferMetadata.log
    ulimit -a >> ${repRID}.inferMetadata.log

    # collect alignment rates (round down to integers)
    align_ercc=\$(echo \$(grep "Overall alignment rate" ERCC.alignSampleSummary.txt | cut -f2 -d ':' | cut -f2 -d ' ' | tr -d '%'))
    align_ercc=\$(echo \${align_ercc%.*})
    echo -e "LOG: alignment rate to ERCC: \${align_ercc}" >> ${repRID}.inferMetadata.log
    align_hu=\$(echo \$(grep "Overall alignment rate" GRCh.alignSampleSummary.txt | cut -f2 -d ':' | cut -f2 -d ' ' | tr -d '%'))
    align_hu=\$(echo \${align_hu%.*})
    echo -e "LOG: alignment rate to GRCh: \${align_hu}" >> ${repRID}.inferMetadata.log
    align_mo=\$(echo \$(grep "Overall alignment rate" GRCm.alignSampleSummary.txt | cut -f2 -d ':' | cut -f2 -d ' ' | tr -d '%'))
    align_mo=\$(echo \${align_mo%.*})
    echo -e "LOG: alignment rate to GRCm: \${align_mo}" >> ${repRID}.inferMetadata.log

    # determine spike-in
    if [ 1 -eq \$(echo \$(expr \${align_ercc} ">=" 10)) ]
    then
      spike="yes"
    else
      spike="no"
    fi
    echo -e "LOG: inference of strandedness results is: \${spike}" >> ${repRID}.inferMetadata.log

    speciesError=false
    speciesError_details=""
    # determine species
    if [ 1 -eq \$(echo \$(expr \${align_hu} ">=" 40)) ] && [ 1 -eq \$(echo \$(expr \${align_mo} "<" 40)) ]
    then
      species="Homo sapiens"
      bam="GRCh.sampled.sorted.bam"
      bed="./GRCh/genome.bed"
      echo -e "LOG: inference of species results in: \${species}" >> ${repRID}.inferMetadata.log
    elif [ 1 -eq \$(echo \$(expr \${align_mo} ">=" 40)) ] && [ 1 -eq \$(echo \$(expr \${align_hu} "<" 40)) ]
    then
      species="Mus musculus"
      bam="GRCm.sampled.sorted.bam"
      bed="./GRCm/genome.bed"
      echo -e "LOG: inference of species results in: \${species}" >> ${repRID}.inferMetadata.log
    else
      echo -e "LOG: ERROR - inference of species returns an ambiguous result: hu=\${align_hu} mo=\${align_mo}" >> ${repRID}.inferMetadata.log
      if [ "${speciesForce}" == "" ]
      then
        speciesError=true
        speciesError_details="**Inference of species returns an ambiguous result:** Percent aligned to human = \${align_hu} and percent aligned to mouse = \${align_mo}"
      fi
    fi
    if [ "${speciesForce}" != "" ]
    then
      speciesError=false
      echo -e "LOG: species overridden to: ${speciesForce}"
      species="${speciesForce}"
      if [ "${speciesForce}" == "Homo sapiens" ]
      then
        bam="GRCh.sampled.sorted.bam"
        bed="./GRCh/genome.bed"
      elif [ "${speciesForce}" == "Mus musculus" ]
      then
        bam="GRCm.sampled.sorted.bam"
        bed="./GRCm/genome.bed"
      fi
    fi

    if [ "\${speciesError}" == false ]
    then
      # infer experimental setting from dedup bam
      echo -e "LOG: infer experimental setting from dedup bam" >> ${repRID}.inferMetadata.log
      infer_experiment.py -r "\${bed}" -i "\${bam}" 1>> ${repRID}.infer_experiment.txt
      echo -e "LOG: inferred" >> ${repRID}.inferMetadata.log

      ended=`bash ${script_inferMeta} endness ${repRID}.infer_experiment.txt`
      fail=`bash ${script_inferMeta} fail ${repRID}.infer_experiment.txt`
      if [ \${ended} == "PairEnd" ]
      then
        ends="pe"
        percentF=`bash ${script_inferMeta} pef ${repRID}.infer_experiment.txt`
        percentR=`bash ${script_inferMeta} per ${repRID}.infer_experiment.txt`
      elif [ \${ended} == "SingleEnd" ]
      then
        ends="se"
        percentF=`bash ${script_inferMeta} sef ${repRID}.infer_experiment.txt`
        percentR=`bash ${script_inferMeta} ser ${repRID}.infer_experiment.txt`
      fi
      echo -e "LOG: percentage reads in the same direction as gene: \${percentF}" >> ${repRID}.inferMetadata.log
      echo -e "LOG: percentage reads in the opposite direction as gene: \${percentR}" >> ${repRID}.inferMetadata.log
      if [ 1 -eq \$(echo \$(expr \${percentF#*.} ">" 2500)) ] && [ 1 -eq \$(echo \$(expr \${percentR#*.} "<" 2500)) ]
      then
        stranded="forward"
      elif [ 1 -eq \$(echo \$(expr \${percentR#*.} ">" 2500)) ] && [ 1 -eq \$(echo \$(expr \${percentF#*.} "<" 2500)) ]
      then
        stranded="reverse"
      else
        stranded="unstranded"
      fi
      echo -e "LOG: stradedness set to: \${stranded}" >> ${repRID}.inferMetadata.log
    else
      ends=""
      stranded=""
      spike=""
      species=""
      percentF=""
      percentR=""
      fail=""
      touch ${repRID}.infer_experiment.txt
    fi

    # write inferred metadata to file
    echo "\${ends},\${stranded},\${spike},\${species},\${align_ercc},\${align_hu},\${align_mo},\${percentF},\${percentR},\${fail}" > infer.csv

    # save species error file
    echo "\${speciesError},\${speciesError_details}" > speciesError.csv
    """
}

// Split metadata into separate channels
endsInfer = Channel.create()
strandedInfer = Channel.create()
spikeInfer = Channel.create()
speciesInfer = Channel.create()
align_erccInfer = Channel.create()
align_huInfer = Channel.create()
align_moInfer = Channel.create()
percentFInfer = Channel.create()
percentRInfer = Channel.create()
failInfer = Channel.create()
inferMetadata_fl.splitCsv(sep: ",", header: false).separate(
  endsInfer,
  strandedInfer,
  spikeInfer,
  speciesInfer,
  align_erccInfer,
  align_huInfer,
  align_moInfer,
  percentFInfer,
  percentRInfer,
  failInfer
)

// Replicate metadata for multiple process inputs
endsInfer.into {
  endsInfer_checkMetadata
  endsInfer_alignData
  endsInfer_countData
  endsInfer_dataQC
  endsInfer_aggrQC
  endsInfer_uploadQC
  endsInfer_failExecutionRun
}
strandedInfer.into {
  strandedInfer_checkMetadata
  strandedInfer_alignData
  strandedInfer_countData
  strandedInfer_aggrQC
  strandedInfer_uploadQC
  strandedInfer_failExecutionRun
}
spikeInfer.into{
  spikeInfer_checkMetadata
  spikeInfer_getRef
  spikeInfer_aggrQC
  spikeInfer_uploadExecutionRun
  spikeInfer_failExecutionRun
}
speciesInfer.into {
  speciesInfer_checkMetadata
  speciesInfer_getRef
  speciesInfer_aggrQC
  speciesInfer_uploadExecutionRun
  speciesInfer_uploadProcessedFile
  speciesInfer_failExecutionRun
}

// Split species count error into separate channel
speciesError = Channel.create()
speciesError_details = Channel.create()
speciesError_fl.splitCsv(sep: ",", header: false).separate(
  speciesError,
  speciesError_details
)

//  Replicate errors for multiple process inputs
speciesError.into {
  speciesError_checkMetadata
  speciesError_uploadExecutionRun
  speciesError_getRef
  speciesError_alignData
  speciesError_dedupData
  speciesError_makeBigWig
  speciesError_countData
  speciesError_fastqc
  speciesError_dataQC
  speciesError_aggrQC
  speciesError_uploadQC
  speciesError_uploadProcessedFile
  speciesError_uploadOutputBag
  speciesError_failPreExecutionRun
}

/* 
 * checkMetadata: checks the submitted metada against inferred
*/
process checkMetadata {
  tag "${repRID}"

  input:
    val endsMeta from endsMeta_checkMetadata
    val strandedMeta from strandedMeta_checkMetadata
    val spikeMeta from spikeMeta_checkMetadata
    val speciesMeta from speciesMeta_checkMetadata
    val endsInfer from endsInfer_checkMetadata
    val strandedInfer from strandedInfer_checkMetadata
    val spikeInfer from spikeInfer_checkMetadata
    val speciesInfer from speciesInfer_checkMetadata
    val fastqCountError_checkMetadata
    val fastqReadError_checkMetadata
    val speciesError_checkMetadata

  output: