blast_template.txt

#!/usr/bin/env bash
#PBS -q omp
#PBS -l mem=@MEM@gb
#PBS -l ncpus=@CORE@
#PBS -l walltime=@WTIME@:00:00

# which version of BLAST do we have to use?
BLAST_VER="2.2.31"

# where do we have to create result files?
WK_DIR=$DATAWORK/blast/${BLAST_VER}

# Type of query: P(rotein) or N(ucleotide) or M(Nucleotide for Megablast)
# Queries below target FASTA formatted sequence files
Q_TYPE=@QTYPE@
case "$Q_TYPE" in
  "P" ) 
    QUERY=/home/ref-bioinfo/beedeem/p/Genbank_Yersinia_Angola_prot/current/Genbank_Yersinia_Angola_prot/Genbank_Yersinia_Angola_prot00
    ;;
  "N" | "M" ) 
    QUERY=/home/ref-bioinfo/beedeem/n/Genbank_Yersinia_Angola_CDS/current/Genbank_Yersinia_Angola_CDS/Genbank_Yersinia_Angola_CDS00
    ;;
  *) 
    echo "ERROR: unknown query type: $Q_TYPE"
    exit 1
    ;;
esac

# Type of subject: P or PS(Protein Small), PL (Protein Large), N or NS (Nucleotide Small) or NL (Nucleotide Large)
S_TYPE=@STYPE@
# Subjects below target BLAST databank alias files (.nal or .pal file)
case "$S_TYPE" in
  "P" | "PS") 
    SUBJECT=/home/ref-bioinfo/beedeem/p/Uniprot_SwissProt/current/Uniprot_SwissProt/Uniprot_SwissProtM
    SUBJECT_NAME=swiss
    ;;  
  "PL")
    SUBJECT=/home/ref-bioinfo/beedeem/p/Uniprot_TrEMBL/current/Uniprot_TrEMBL/Uniprot_TrEMBLM
    SUBJECT_NAME=trembl
    ;;  
  "N" | "NS") 
    SUBJECT=/home/ref-bioinfo/beedeem/n/Genbank_Bacteria/current/Genbank_Bacteria/Genbank_BacteriaM
    SUBJECT_NAME=gb_bact
    ;; 
  "NL")
    SUBJECT=/home/ref-bioinfo/beedeem/n/NCBI_nt/current/NCBI_nt/NCBI_ntM
    SUBJECT_NAME=nt
    ;;  
  *)  
    echo "ERROR: unknown subject type: $S_TYPE"
    exit 1
    ;;  
esac

# BLAST output format
#  5:XML ; 9: binary asn1 ; 6: tabular ; see: https://www.ncbi.nlm.nih.gov/books/NBK279684/ : Table C1
#  we use ASN.1 since it can be used to generate other ones using blast_formatter tool
FORMAT=9

case "$FORMAT" in
  "5" )
    FORMAT_EXT="xml"
    ;;
  "6" )
    FORMAT_EXT="tab"
    ;;
  "9" ) 
    FORMAT_EXT="asn1"
    ;;
  *) 
    FORMAT_EXT="out"
    ;;
esac

# BLAST command to use
BLAST_CMD="?"
QT=${Q_TYPE:0:1}
ST=${S_TYPE:0:1}
if [ "$QT" == "P"  ] && [ "$ST" == "P"  ]; then
  BLAST_CMD="blastp"
  BLAST_CMD_NAME="blastp"
elif [ "$QT" == "N"  ] && [ "$ST" == "P"  ]; then
  BLAST_CMD="blastx"
  BLAST_CMD_NAME="blastx"
elif [ "$QT" == "M"  ] && [ "$ST" == "N"  ]; then
  # Megablast
  BLAST_CMD="blastn"
  BLAST_CMD_NAME="blastn"
elif [ "$QT" == "N"  ] && [ "$ST" == "N"  ]; then
  # blastn
  BLAST_CMD="blastn -task blastn"
  BLAST_CMD_NAME="megablast"
fi

# Output file name
OUTFNAME=$WK_DIR/${Q_TYPE}-${S_TYPE}-@MEM@gb-@CORE@c
TIMEFNAME=${OUTFNAME}.time
OUTFNAME=${OUTFNAME}.${FORMAT_EXT}

# Prepare a TIME specific format string
export TIME="U=%U;S=%S;E=%E;P=%P;M=%M;K=%K;F=%F;W=%W;I=%I;O=%O"

# Load BLAST Conda environment
. /appli/bioinfo/blast/${BLAST_VER}/env.sh

# Prepare BLAST command
CMD="/usr/bin/time -o $TIMEFNAME --format='$TIME' $BLAST_CMD -query $QUERY -db $SUBJECT -out $OUTFNAME -outfmt $FORMAT -max_target_seqs 1 -evalue 1e-3 -num_threads @CORE@"
# dump cmd to be executed in PDB log for this job
echo $CMD
# eventually, run BLAST!
eval $CMD