Creating a loop to mark duplicates using Picard

Question

[brandnewatthis]

Hi there,

I am trying to create a loop to run multiple sorted bam files through picard to mark duplicates. My bash script is not working, and I'm at a loss.

#! /bin/bash

picard=/home/picard/picard.jar
sbam=/data/B_bam

for bamfile in *_sorted.bam
do
    java -jar "$picard" MarkDuplicates INPUT="$sbam"/"$bamfile" OUTPUT="$sbam"/"${bamfile%_sorted.bam}"_md.bam METRICS_FILE="$sbam"/"${sbam%_sorted.bam}"_metrix
    samtools index "$sbam"/"${bamfile%_sorted.bam}"_md.bam
done

Error Message:

Cannot read non-existent file: /data/B_bam/*_sorted.bam

Answer 1

[Pierre Lindenbaum]

METRICS_FILE="$sbam"/"${bamfile%_sorted.bam}"_metrix

instead of

METRICS_FILE="$sbam"/"${sbam%_sorted.bam}"_metrix

also check there is a bam file under in the CURRENT WORKING directory: for bamfile in *_sorted.bam but may be you wanted for bamfile in /data/B_bam/*_sorted.bam

also don't use a loop in a script shell when you can use Snakemake or Nextflow

[Ram]

don't use a loop in a script shell when you can use Snakemake or Nextflow

People need to start somewhere. Loops -> xargs/parallel -> snakemake/nextflow is the logical progression in learning automation IMO.

I find it weird that OP uses a mix of ${} and "..."/"...." (multiple quoted strings interleaved with multiple unquoted strings). Why not simply "${sbam}/${bamfile%_sorted.bam}_metrix"?

[brandnewatthis]

Sorry, I am very new at this and am just learning my way around these symbols! Thanks for the help :)

[Ram]

For someone new at shell scripts, you're doing really well at using shell variables and parameter expansion. Good going!