This site is now read-only. You can find our new documentation site and support forum for posting questions here.
Be sure to read our welcome blog!
MarkDuplicates out of Memory
I have been using Picard MarkDuplicates function to mark the duplicate reads which is found to be utlizing huge amount of memory compared to the memory used by GATK in subsequent steps.
I have used the below command:
java -Xmx128G -jar /appl/bio/picard/picard-2.6.0/picard/build/libs/picard.jar MarkDuplicates INPUT=alignment_sorted.bam OUTPUT=alignment_duprem.bam CREATE_INDEX=true VALIDATION_STRINGENCY=LENIENT QUIET=true AS=true METRICS_FILE=metrics VERBOSITY=WARNING
which utilized ~50GB to complete the run. I tried to minimize the memory consumption by adding MAX_RECORDS_IN_RAM=50000 to the above command but it utilized the same amount of memory as before.
I tried searching different forums but did not find a solution for this or probably i am missing some discussions. Could someone throw some light into this issue to decrease the memory consumption.