run.sh

#!/bin/bash

#SBATCH --time=01:00:00   # walltime
#SBATCH --ntasks=1   # number of processor cores (i.e. tasks)
#SBATCH --nodes=1   # number of nodes
#SBATCH --mem-per-cpu=64G   # memory per CPU core
#SBATCH -J "PepseqPipeline"   # job name

# LOAD MODULES, INSERT CODE, AND RUN YOUR PROGRAMS HERE
module load r/3/3
module load jdk/1.8.0-121




# exit if any of the commands fail
set -e

#Set clean up function to be called if error occurs and program exits
#Removes all the temp files created in the bash script
function cleanUp {
	rm -rf temp/
	rm -f Rplots.pdf
	echo "ERROR: pep-seq pipeline script failed"
}
trap cleanUp ERR

#Set usage 
USAGE="
USAGE:
./run.sh [input_file_name] [-options]

OPTIONS:
	--arff: data already in arff format, don't convert data
	--anti: also find antitoxic motifs
	--neutral: also find neutral motifs
	--help: print usage
	-b: balance the data before running machine learning classifier
	-k: [number_of_motifs]: specify the number of motifs to find
	-o: [out_dir]: specify directory in results/ in which to save the output files
	-a: print ALL toxic motifs
"

#Check to see if there is at least one command line argument
if [ $# -lt 1 ]
then
	echo "ERROR: You must specify input file as command line argument"
	echo "$USAGE"
	exit 1
fi

PEP_LIBRARY=$1
INPUT_FILE=$PEP_LIBRARY
shift

#Check if Input file is a valid file
if [ ! -f $INPUT_FILE ]
then
	echo "ERROR: Input file $1 is not valid"
	echo "$USAGE"
	exit 1
fi

#Initialize parameter values
K=30
arff=false
anti=false
neutral=false
balance=false
output=false
all=false

#Iterate through command line arguments and store data into variables
while [ $# -gt 0 ]
do
	ARG="$1"
	case "$ARG" in
		--arff) #data already in arff format
			arff=true
			;;
		-k) #Specify number of motifs to test (default is 30)
			K=$2
			shift
			;;
		--anti) #Also find antitoxic motifs
			anti=true
			;;
		--neutral) #Also find neutral motifs
			neutral=true
			;;
		-b) #Balance the data before running random forest classifier
			balance=true
			;;
		-o) #Save output of the model to a new direcrtory in results
			output=true
			OUTDIR=$2
			shift
			;;
		-a) #find all toxic motifs from random forest
			all=true
			;;
		--help) #Print Usage message
			echo "Help Requested."
			echo "$USAGE"
			exit 1
			;;
		*) #Default branch, print argument not valid and usage then exit
			echo "Not a valid argument: $1"
			echo "$USAGE"
			exit 1
			;;
	esac
	shift
done

MotifFinderParam="-k $K"
if [ $all = true ]
then
	MotifFinderParam="-a"
fi 


#make temp directory to store temporary files
mkdir -p temp

#IF balance parameter was passed in, run python balance script on the input data
if [ $balance = true ] && [  ! $arff = true ]
then
	>&2 echo "Balancing input data . . . "
	python oversampling.py $INPUT temp/balancedtemp.csv
	INPUT_FILE=temp/balancedtemp.csv
fi

#If passed in input file is not already an arff file, convert it to arff format
if [ ! $arff = true ]
then
	>&2 echo "Converting input data to an arff file . . ."
	cat $INPUT_FILE | python py_scripts/convert_to_arff.py > temp/arfftemp.arff
	INPUT_FILE=temp/arfftemp.arff
	rm -f temp/balancedtemp.csv
fi

#Run Weka's random forest classifiers on the arff file and store the tree output into temp.txt
>&2 echo "Running Random Forest classification on data . . . "

#module load jdk/1.8.0-121 #(Uncomment this line if java not updated)

java -Xmx32g -cp dependency_jars/weka.jar weka.classifiers.trees.RandomForest -U -B -V 1e-6 -P 30 -I 500 -no-cv -print -t $INPUT_FILE > temp/foresttemp.txt
rm -f temp/arfftemp.arff

MOTIF_FILE=temp/motifs.txt
#Take the output of weka's Random Forest classifier and put it into our MotifFounder Algorithm
>&2 echo "Extracting motifs from Random Forest . . . "
if [ $anti = true ] && [ $neutral = true ]
then
	java -jar dependency_jars/MotifFinder.jar temp/foresttemp.txt $MotifFinderParam > temp/RFmotifs.txt
elif [ $anti = true ]
then
	java -jar dependency_jars/MotifFinder.jar temp/foresttemp.txt $MotifFinderParam -noneu > temp/RFmotifs.txt
elif [ $neutral = true ]
then
	java -jar dependency_jars/MotifFinder.jar temp/foresttemp.txt $MotifFinderParam -noanti > temp/RFmotifs.txt
else
	java -jar dependency_jars/MotifFinder.jar temp/foresttemp.txt $MotifFinderParam -noneu -noanti > temp/RFmotifs.txt
fi
rm -f temp/foresttemp.txt
sed 's/[()]//g' temp/RFmotifs.txt > temp/rfmotifs.txt
sed 's/\//\t/g' temp/rfmotifs.txt > temp/RFmotifs.txt
rm -rf temp/rfmotifs.txt

#Calculate motif coverage and motif accuracy of the selected motifs and print
#these values out in the motif file that was created
>&2 echo "Creating Motif Set from extracted motifs . . ."
echo "$1"
./cppScripts/build/scoreMotifs $PEP_LIBRARY temp/RFmotifs.txt > temp/motifs.csv 2> temp/statistics.txt
mv clusteredPeps.csv temp/
rm -f temp/RFmotifs.txt

>&2 echo "Testing significance of individual motifs . . ."
##module load r/3/3 #(Include if R module not loaded)
Rscript --vanilla R_scripts/chi_squared.R temp/motifs.csv &> /dev/null

>&2 echo "Calculating statistics on the Motif Set . . .  "
##Run motifSet T test for peps inside and outside of the motif set
Rscript --vanilla R_scripts/motifSetTtest.R temp/clusteredPeps.csv >> temp/statistics.txt 2> /dev/null
rm -f temp/clusteredPeps.csv

#If output is true, save the output file to the specifies directory in results. If it does not exist, create such a directory
#If output not specified print out the motifs data to standard output
if [ $output = true ]
then
	>&2 echo "Saving output to results/$OUTDIR . . . "
	mkdir -p results/$OUTDIR
	mv temp/motifs.csv results/$OUTDIR
	mv temp/statistics.txt results/$OUTDIR
	mv temp/motifSetBoxPlot.png results/$OUTDIR
else
	cat temp/motifs.csv
	echo
	cat temp/statistics.txt
fi

rm -rf temp/
>&2 echo "Pep-seq pipeline executed successfully!"
exit 0