BWA

#!/bin/bash -x
#PBS -N bwa_job
#PBS -q batch
#PBS -l walltime=72:00:00
#PBS -l feature=fastest_nodes_on_your_server
#PBS -l nodes=1:ppn=4
#PBS -j oe
#PBS -A your_job

#bwa script for more than two reads per sample, that needs to be merged or piled after
INPATH=/mnt/home/infiles
OUTPATH=/mnt/home/outfiles
mkdir -p $OUTPATH

reference=/mnt/home/user/references/refseq.fasta
cd /mnt/home/usr/tools/bwa-0.7.17
./bwa index $reference

shopt -s nullglob
for dir in $INPATH{/,/*} ;
do
        echo dir is $dir
        for file in $dir/*.gz ;
        do
                #define sample ID for including in the read group
                ID=$(basename $dir)
                echo ID is $ID
                bname=$(basename $file )
                sample=$( cut -d'_' -f 1,2,3<<<"$bname")
                echo bname sample is $bname $sample
                input1=$dir/$sample"_1.paired.fq.gz"
                input2=$dir/$sample"_2.paired.fq.gz"
                outfile=$OUTPATH/$ID".sam"
                #check sample names before pairing
                echo input1 is  $input1 output1 is $output1 input2 is $input2 output $output2
                /mnt/home/usr/tools/bwa-0.7.17 mem -R "@RG\tID:1\tLB:1\tPL:MGI\tPU:1\tSM:$ID" $reference $input1 $input2 > $outfile
                echo $outfile
        done
done