nextflow.config

manifest {
    name = 'CompGenLabUB/CAPTVRED'
    author = 'Maria Tarradas-Alemany <mtarradasi@ub.edu>;Josep F Abril Ferrando <jabril@ub.edu>'
    description = 'Pipeline for analyses of viral metagenomics datasets from Target Enrichment Sequencing.'
    version = '1.0.0'
    mainScript = 'main.nf'
    nextflowVersion = '>=22.04.5'
}

params{
  //environment parameters:
    //help
    help = false

    conda.enabled = true
    
    // init params
    // working directory
    samp        =  null  // Sample definition tabular filef
    fastq_dir   =  null  // Directory for target species data
  
    runID       =  null  // "RUN00"
    R1          =  "_R1_001"
    R2          =  "_R2_001"
    setname     =  "targetset"

    // refseqs and databases
    refdb_name    =  "rvdb_nt"
    set_seqs      =  "setseqs.fasta.gz"  // Fasta file for set species is stores (full path)
    fams_subset   =  "foi_subset_db.fasta.gz" //subset of rvdb conaining only 
    others_subset =  "other_subset_db.fasta.gz"
    full_tax      =  "full_tax.tax.gz"
    set_tax       =  "set_tax.tax.gz"
    use_full_db   =  params.use_full_db?:false

    amplicon_refseqs_dir   =  null //params.refseqs+'/ampliconseqs'
    amplicon_refseqs_info  =  "Viral_candidates_zoonosis_full_info.tsv"

    rawfq_sfx="{$R1,$R2}.{fastq,fq}.gz"
    // sampletbl= params.basedir+'/samples_definition.tbl'
    
  //
    NCPUS=32
    // MAXMEM="$MAXMEM"
    // phred="$PhredOffset"up
    
    
  // Cleaning params (BBDuk):
    trim_adapters = true
   // bbdukREF="${params.refseqs}/bbmap/resources/adapters.fa"
    bbdukMINLEN=32
    bbdukMAQ=10
    
  // BOWTIE and Alignment params:
    alignMINQ=13
    bowtie_nmismatch=1
    bowtie_seedlen=18
    
  // Assembly params:
    assembler="megahit"
    assemblyMINCONLEN=100;
    
    trinityMAXM="250G"
    megahit_errorHandler = "finish"

  
  // Taxonomy
     //taxonfast = false
     //taxonslow = false
    taxalg="BLASTN"
    
  // Blast params:
    // blast_approach="BLASTN"
    // blast_ref_db_name="C-RVDBvCurrent_blastdb"
    // blast_ref_db_name="$BL_DB_NM" 
       //"C-RVDBvCurrent_blastdb"
    // blast_ref_db_ids="$BDBD_IDS"
   //  blast_ref_db_info="C-RVDB_allentries_info.txt"
    blast_eval=10e-10
    blast_pident=50
    taxondbsize=5000000000 // about 2 times db size (2428170463)
    
  // ON REFERENCES (SET)
   // general_only = false
  
  // Blast against Viral Candidates (vcan)
    bl_suffix="blastn_on_viralcandidates"
    vcan_eval=0.001
    vcan_pident=50
    
    // do not modify! //
    bl_outfmt="6 qseqid qlen sseqid slen qstart qend sstart send"
    bl_outfmt="${bl_outfmt} length score evalue bitscore pident"
    bl_outfmt="${bl_outfmt} nident mismatch positive gapopen gaps"
    bl_outfmt="${bl_outfmt} ppos qframe sframe qcovs qcovhsp qseq sseq"
    // do not modify! //
    
  // Kaiju Params:
    kaijumaxCPUS=26 
    kaijuMAX_FORKS=4
    kaijuMAX_RETRIES=2
    kaijuDBRAW="nr_euk"   // database for kaiju into raw reads

  // Filter params:
    mincovpct=70

  // Handle Contamination:
    handle_contamination = params.handle_contamination?:false
    cont_min_cov=100;
  //  if (handle_contamination == true ) {
  //    cids="$contamin_ids";
  //    cfaa="$contamin_faa";
  //  }

  //Coverage figures
  do_cov_figures == true
    
}

process {
    
    withName: '*' { 
       conda = '$baseDir/environment.yml' 
    }
    
    withLabel: 'trinity_crash' {
       errorStrategy = 'ignore'
    }
    
    withLabel: 'megahit_crash' {
       errorStrategy = params.megahit_errorHandler
    }

    withLabel: 'limit_kaiju' {
       errorStrategy = 'retry'
       maxRetries = params.kaijuMAX_RETRIES
       maxForks = params.kaijuMAX_FORKS
       if ( params.NCPUS < params.kaijumaxCPUS ) { 
            cpus=params.NCPUS 
       } else { 
            cpus=params.kaijumaxCPUS  }
              
       
    }
}

env {

    //Fixing issues with python utf8 libraries
    LANG = "en_US.UTF-8"
    LC_CTYPE = "en_US.UTF-8"
    LC_NUMERIC = "en_US.UTF-8"
    LC_TIME = "en_US.UTF-8"
    LC_COLLATE = "en_US.UTF-8"
    LC_MONETARY = "en_US.UTF-8"
    LC_MESSAGES = "en_US.UTF-8"
    LC_ALL = "en_US.UTF-8"
    
      // Kaiju params
    //kaiju_dbs="viruses, rvdb, refseq, nr_euk"
    
    
    //Further defs
    //BP_LOG = "$BDIR/logs/biopieces"
    BPCOL = 'SEQ_NAME,SEQ_LEN,RES[A],RES[C],RES[G],RES[T],GC%'
    BPCOL = "$BPCOL,SCORES_MIN,SCORES_MAX,SCORES_MEAN,SCORES_MEDIAN"

    TAB = "\t"
    RET = "\n"

    ZERO="###0###"
}