sge_hmm.sh

#!/bin/bash -l
#$ -cwd
#$ -j y
#$ -t 1-10:1
#$ -tc 25
#$ -N hmm_unk
#$ -pe threaded 2
#$ -V

# Where the models are
DIR=/bioinf/projects/megx/UNKNOWNS/chiara/HMP

# Choose models based on the task ids inside the file pfam_hmm_keys
HMM=$(awk -vL="${SGE_TASK_ID}" "NR == L" ${DIR}/pfam_hmm_keys)

# Where the models are
PFAM="${DIR}"/database

# Out dir
ODIR=/bioinf/home/afernand/SANDBOX/hmm_unknwown_test

# Fasta file to annotate distributed in the cluster nodes
INFILE=/scratch/hmm_unknowns/TARA_OSD_GOS_malaspina.orf.aa.renamed.fasta

# Where the results and logs are going to be saved
H_RES="${ODIR}"/results

# The search... NSLOTS is a special SGE variable related to the Parallel
# Environment (-pe threaded 2)
# --cut_ga is related to the hmmer search, not used in our case as we don't have
# this threshold defined
# -Z Search space... (# queries * # models) or (# models * # queries)
# --domtblout output of the search based on the domains, check HMMER docs
# -o log file
# "${PFAM}"/pfam_HMM."${HMM}" the model
# "${INFILE}" Fasta file
hmmsearch --cpu "${NSLOTS}" \
  --cut_ga \
  -Z 159561257 \
  --domtblout "${H_RES}"/HMPPfam30."${HMM}".out \
  -o "${H_RES}"/HMPPfam30."${HMM}".log \
  "${PFAM}"/pfam_HMM."${HMM}" \
  "${INFILE}"