Remove flags of MarkDuplicates (picard)

Question

[Coryza]

Hi,

Is it possible to remove the MarkDuplicates flags (not the sequences) from a BAM file? If so, how?

[Pierre Lindenbaum]

duplicate of Tool to unmark duplicates

Answer 1

[Malachi Griffith]

You can use Picard RevertSam for this. This tool can be used to reset various attributes of a BAM file including duplicate information. Simply use: REMOVE_DUPLICATE_INFORMATION=true

Example command:

java -Xmx7g -jar ~/tools/picard/picard-tools-1.118/RevertSam.jar OUTPUT=UnmarkedDuplicates.bam INPUT=MarkedDuplicates.bam REMOVE_DUPLICATE_INFORMATION=true

[Pierre Lindenbaum]

RevertSam is new to me. Thanks

Answer 2

[Devon Ryan]

Depending on the version of awk you have on your computer then something like the following should work:

samtools view -h foo.bam | awk 'BEGIN{OFS="\t"}{if(NF>5) {if(and($2,1024)) {$2-=1024}} print $0}' | samtools view -Sbo foo.unmarked.bam -

I think Macs have mawk rather than gawk, so this doesn't work there.

[Ram]

If you're on Mac and not using homebrew, you're missing out on a bunch of cool stuff.

[Coryza]

Thanks! Worked perfectly ;)

[Neilfws]

Worked for me but the next stage of my pipeline (realignment using GATK) did not like the resulting BAM files.

##### ERROR MESSAGE: SAM/BAM file /home/cci/sau103/datahome/sals/scratch/Run/bowtie2/100028_S233_L001.asd.bam is malformed: Invalid file pointer: 4570