#!/bin/bash -l

#Specify project
#$ -P scv #replace with your own project_name

#Give the name to the job
#$ -N samtools_example

#Send an email when the job is finished (or aborted)
#$ -m ae

#Join the error and output file
#$ -j y

#Specify qlog directory: 
#$ -o ../qlog

#Request multiple cores:
#since samtools support multithreading, we request 2 cores in this example 
#$ -pe omp 2

# Now let's keep track of key job information:
echo "=========================================================="
echo "Starting on : $(date)"
echo "Running on node : $(hostname)"
echo "Current directory : $(pwd)"
echo "Current job ID : $JOB_ID"
echo "Current job name : $JOB_NAME"
echo "Number of cores: $NSLOTS"
echo "=========================================================="


# define variables: 
INPUT_DIR=../sam_in
OUT_DIR=../sam_out

#Sepcify the version of samtools
module load samtools/1.10

# run samtool view to convert sam to bam
# and pipe to samtool sort to get the sorted bam
# use $NSLOTS to make sure the command only use the number of cores requested
# incorporate JOB_NAME in the output file name to distinguish output accordingly
# assign command string to CMD variable
CMD="samtools view --threads $NSLOTS -bSu ${INPUT_DIR}/example_aln.sam | samtools sort --threads $NSLOTS -o ${OUT_DIR}/${JOB_NAME}.sorted.bam"
echo $CMD
eval $CMD

# print out end message: 
echo "DONE!"