pan_pipe

#!/bin/bash
set -e

scriptName="pan_pipe"
version="0.0.1"

## download and collect all of the genomes you want in your pangenome
# it is very important to QC this collection.
# often genomes are mislabeled in genbank or of poor quality



HELP=0
THREADS=1
ROARY_ARGS=''
PROKKA_ARGS=''
GIFROP_ARGS=''

# parse arguments here

for arg in "$@"
do
    case $arg in
	-h|--help)
	HELP=1
	shift
	;;
        --roary_args)
        ROARY_ARGS=$2
        shift # Remove --initialize from processing
        ;;
        --prokka_args)
        PROKKA_ARGS=$2
        shift # Remove --initialize from processing
        ;;
        --gifrop_args)
        GIFROP_ARGS=$2
        shift # Remove argument name from processing
        # shift # Remove argument value from processing
        ;;
        -t|--threads)
        THREADS="$2"
        shift # Remove argument name from processing
        # shift # Remove argument value from processing
        ;;
        *)
        OTHER_ARGUMENTS+=("$1") # arg values seem to be accumulating in here
        shift # Remove generic argument from processing
        ;;
    esac
done


# help message here

# Print usage
usage() {
  echo -n "

Usage:

  ${scriptName} [OPTION]...

This script should be executed from a directory that contains bacterial genomes in fasta format
with the suffix ".fna".  This script will annotate all genomes with prokka, then calculate a
pangenome with roary, finally it will extract genomic islands from the pangenome with gifrop.
Options can be passed to prokka, roary, and gifrop.

IMPORTANT!!

Inputs should be carefully QC'd.  Garbage in, garbage out.

 Options:
  -h, --help        Display this help and exit
  -t, --threads     Number of threads to use for parallel commands. Will be overridden by values in *_args options
  --roary_args      a quoted string of arguments to pass to roary, e.g.: '-p 8 -s -e --mafft'
  --prokka_args     a quoted string of arguments to pass to prokka, e.g: '--rawproduct --proteins prots.gbk'
  --gifrop_args     a quoted string of arguments to pass to gifrop, e.g: '--no_plots -m 7'

Example:

pan_pipe
or
pan_pipe --gifrop_args '--min_genes 5 --no_plots' --roary_args '-s' --prokka_args '--proteins prots.gbk'


"
}


if [[ $HELP -eq 1 ]]
 then
     usage
     exit
fi



# check deps

# check dependencies
# ERR=0
echo "===== Dependencies check ====="
[ -z `which parallel` ]      && echo "parallel      .... not found" && ERR=1 || echo "parallel      .... good"
[ -z `which abricate` ]      && echo "abricate      .... not found" && ERR=1 || echo "abricate      .... good"
[ -z `which Rscript` ]       && echo "Rscript       .... not found" && ERR=1 || echo "Rscript       .... good"
[ -z `which find` ]          && echo "find          .... not found" && ERR=1 || echo "find          .... good"
[ -z `which prokka` ]        && echo "prokka        .... not found" && ERR=1 || echo "prokka        .... good"
[ -z `which roary` ]         && echo "roary         .... not found" && ERR=1 || echo "roary         .... good"
[ -z `which gifrop` ]        && echo "girfop        .... not found" && ERR=1 || echo "gifrop        .... good"

if [[ $ERR -eq 1 ]]
 then
     echo "Link or install any of your 'need to install' programs above"
     exit
fi


### CHECK FOR '.fna' FILES HERE: ###


numFNA=$(find . -name '*.fna' | wc -l)

if [[ $numFNA -lt 2 ]]
 then
     echo "less than two fastas detected. Cannot calculate pangenome with less than two fastas"
     echo "Please make sure your fasta files end in '.fna'"

     exit
fi




##### annotate all fasta formatted genomes with prokka #####
# technically this will give an error if there are enough *fna files to give an arguments too long error
# but if you are really trying to make a pangenome with that many fastas, this error will be the least of your worries

mkdir panpipe_logs

echo 'generating prokka commands... see prokka_cmds.txt'

for x in *fna
do
printf 'prokka --prefix %s %s %s\n' "${x%.fna}" "$PROKKA_ARGS" "$x" >> prokka_cmds.txt
done

echo 'done generating prokka commands'

echo 'executing prokka commands in parallel'

echo 'this can take a bit...'
cat prokka_cmds.txt | parallel >> ./panpipe_logs/prokka_logs.txt 2>&1

echo 'done with prokka'
# make a directory for the pangenome analysis
mkdir pan

# then copy all gffs from prokka directories

find . -maxdepth 2 -path ./pan -prune -o -name '*.gff' -exec cp {} ./pan/ \;


# enter the pan directory
cd pan

# run roary

roar_cmd="roary $ROARY_ARGS *gff"
echo "running command: $roar_cmd"

roary $ROARY_ARGS *gff 2>&1 | tee -a ../panpipe_logs/roary.log

echo 'done running roary'

# then run gifrop
echo 'running gifrop'
gif_cmd="gifrop $GIFROP_ARGS --get_islands"
echo "running command: $gif_cmd"

gifrop $GIFROP_ARGS --get_islands 2>&1 | tee -a ../panpipe_logs/gifrop.log

echo 'Done!'
# boom!  now you have a pan genome and genomic islands in the context of this pan genome