3

I have a script the aim of which is:

  • For a list of files, obtain a specific number for each file (this is sequencing data, to be specific) and store these into array1
  • Using array1, find the smallest number is array1
  • Based on smallest number in array1, divide all it by all numbers in array1 to make array2.

My script looks as follows:

#!/usr/bin/bash



USAGE() { echo "Usage: bash $0 [-b <in-bam-files-dir>] [-o <out-dir>] [-c <chromlen>]" 1>&2; exit 1; }

if (($# == 0))
then
    USAGE
fi



while getopts ":b:o:c:h" opt
do
    case $opt in
        b ) BAMFILES=$OPTARG
        ;;
        o ) OUTDIR=$OPTARG
        ;;
        c ) CHROMLEN=$OPTARG
        ;;
        h ) USAGE
        ;;
        \? ) echo "Invalid option: -$OPTARG exiting" >&2
        exit
        ;;
        : ) echo "Option -$OPTARG requires an argument" >&2
        exit
        ;;
    esac
done



if [ ! -d ${OUTDIR} ]
then
    mkdir ${OUTDIR}
fi

if [ ! -d ${OUTDIR}/temp ]
then
    mkdir ${OUTDIR}/temp
fi

if [ -d ${BAMFILES} ]
then
    echo -e "\nProcessing BAM files from following directory: ${BAMFILES} \n "
fi



module purge
module load samtools
module load bedtools
module load ucsctools
echo -e "Modules are loaded\n"



FIRSTBAM=$(ls $BAMFILES/*bam | head -1)
MIN=$(samtools view -c -F 260 ${FIRSTBAM} )
echo -e "Minimum number of reads is currently set to $MIN from $FIRSTBAM (first bam in directory)\n"



declare -A BAMREADS
echo "BAMREADS array is initialized"

for i in $(ls $BAMFILES/*bam)
do
    echo "Counting reads in $i "
    BAMREADS[$i]=$(samtools view -c -F 260 $i)
done



for i in ${BAMREADS[@]}
do
    if [[ $i -lt $MIN ]]
    then
        MIN=$i
    fi
done

echo -e "Minimum number of reads that will be used for scaling is $MIN \n"



declare -A BAMFRACS
echo -e "BAMFRACS array is initialized"

for i in ${!BAMREADS[@]}
do
    BAMFRACS[$i]=$(awk -v var1=${MIN} -v var2=${BAMREADS[$i]} 'BEGIN { x= var1 / var2; printf "%.8f", x }')
done



for i in $(ls $BAMFILES/*bam)
do

    SAMPLE=`basename $i`
    SAMPLE=${SAMPLE%.bam}
    echo $SAMPLE

    if [[ ${BAMREADS[$i]} -eq $MIN ]]
    then

        echo "Sample $i does not need scaling"

        command="cp $i ${OUTDIR}/temp/${SAMPLE}.scaled.bam;
        genomeCoverageBed -bg -split -ibam ${OUTDIR}/temp/${SAMPLE}.scaled.bam > ${OUTDIR}/temp/${SAMPLE}.bedgraph;
        sed -e 's/^/chr/g;s/MT/M/g' ${OUTDIR}/temp/${SAMPLE}.bedgraph > ${OUTDIR}/temp/${SAMPLE}.modified.bedgraph;
        sort -k1,1 -k2,2n ${OUTDIR}/temp/${SAMPLE}.modified.bedgraph > ${OUTDIR}/temp/${SAMPLE}.sorted.bedgraph;
        bedGraphToBigWig ${OUTDIR}/temp/${SAMPLE}.sorted.bedgraph $CHROMLEN ${OUTDIR}/${SAMPLE}.bw"
        #rm ${OUTDIR}/temp/${SAMPLE}.*

    else

        command="samtools view -s ${BAMFRACS[$i]} -b $i > ${OUTDIR}/temp/${SAMPLE}.scaled.bam;
        genomeCoverageBed -bg -split -ibam ${OUTDIR}/temp/${SAMPLE}.scaled.bam > ${OUTDIR}/temp/${SAMPLE}.bedgraph;
        sed -e 's/^/chr/g;s/MT/M/g' ${OUTDIR}/temp/${SAMPLE}.bedgraph > ${OUTDIR}/temp/${SAMPLE}.modified.bedgraph;
        sort -k1,1 -k2,2n ${OUTDIR}/temp/${SAMPLE}.modified.bedgraph > ${OUTDIR}/temp/${SAMPLE}.sorted.bedgraph;
        bedGraphToBigWig ${OUTDIR}/temp/${SAMPLE}.sorted.bedgraph $CHROMLEN ${OUTDIR}/${SAMPLE}.bw"
        #rm ${OUTDIR}/temp/${SAMPLE}.*

    fi

    echo $command | qsub -V -cwd -o $OUTDIR -e $OUTDIR -l tmem=10G -l h_vmem=10G -l h_rt=3600 -N bigwig_${SAMPLE}

 done

 echo "Task completed: conversion jobs submitted to cluster"

I have 2 questions:

  • From what I understand, bash is not very good at doing arithmetic maths: i.e. doing any kind of operation (addition, division etc) involving float numbers. However, given the fact that var1 and var2 are always integers in my script (see $MIN and all array1 values), do we agree that this is not a problem? I.e. my operation results in float numbers, but it uses integers, so it's not a problem right?

  • It's not very clear in StackExchange because there is no syntax highlighting here, but I noticed that the var2=${BAMREADS[$i]} part of my script isn't quite right. I use nano and in my terminal, instead of having all of the ${BAMREADS[$i]} in red, like the other variables (like ${MIN}), only the ${BAMREADS[$i part of the script is appearing in red, i.e. the ending ]} are not red. The script seems to be behaving as I expect and everything seems to be working.. So I don't quite understand why it's not all in red.

This is how my script looks like in nano (notice how the ]} in ${BAMREADS[$i]} in the awk command AND later on in the second $command is not in red as it should be):

enter image description here

However, if you paste this code into https://www.shellcheck.net/, you don't get any problem in terms of highlighting in this part of the script. So how come nano and shellcheck are not telling me the same thing? I have used this script and it seems to work for me but I am concerned by this highlighting issue..

Thanks

4
  • Well it seems that adding double quotes in the awk command somehow solves the problem but I'm quite confused how my script managed to work before. I've testing it on real data and it worked..
    – mf94
    Commented Sep 17, 2018 at 10:07
  • Frankly I'm quite confused. Nano and shellcheck.net don't seem to highlight things in the same ways.
    – mf94
    Commented Sep 17, 2018 at 10:11
  • If you managed to solve the problem, remember to please post an answer and later accept it.
    – simlev
    Commented Sep 17, 2018 at 10:22
  • Just update the question with my 'real' script and more explanations. I am still cofnused by this syntax highlighting conflict between nano and shellcheck.net
    – mf94
    Commented Sep 17, 2018 at 10:50

1 Answer 1

3

Syntax highlighting is one problem

Each editor has its own dis-/advantages in this manner.

See my question on SoftwareRecs and its respective answers, most importantly this for both CLI and GUI and this for GUI.

Notably, Visual Studio Code, has IMHO the best Syntax highlighting from GUI editors.

From CLI editors, refer to the answer of gVim, which does the same syntax highlighting job for CLI.

Note that as I was a heavy nano user, I can tell you nano can't distinguish the variables inside quotes.


Missing double quotes are a bigger problem

What should trouble you most is that you did not use - I suppose you're not used to it - bad habit - the double quotes. Please refer to StackOverflow for more information or simply use Google. Or see below.


Double quote to prevent globbing and word splitting

For shell script writers, nano editor is hardly usable, because it won't recognize variables from inside string (quotes), which is very bad for every shell script writer. Double quotes are completely essential in shell scripts. They prevent so called globbing and word splitting, read ShellCheck Wiki article SC2086 for more information about this topic.


Always pipe your scripts to ShellCheck

1
  • So are you saying essentially that nano is perhaps not the best for syntax highlighting? I read your post, seems like Visual Studio Code is better then. I used ShellCheck and it definitely does not highlight an issue in the place I mentioned in my post so I am going to go ahead and assume it's all good. My script has run successfully before. But as you said, I will make sure I clean up the double quotes. I agree about the double quotes, I'm still quite new to bash scripting so will make sure I clean it up. Thanks!
    – mf94
    Commented Sep 17, 2018 at 12:57

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .