picard MarkDuplicates backsplice reads

Picard does not seem to be removing duplicate backsplice reads.
Did a test run on RNA-seq data (paired end) not using a duplicate removal tool and with picard. The number of unmapped reads (Bowtie2 from the find_circ pipeline) is the same for both methods. Even after a second round of duplicate removal with picard on the unmapped reads bam-file, the number of reads remains the same.

Used versions and commands:
picard 2.1.1
Java 1.8.0_74

java -jar picard.jar MarkDuplicates I=accepted_hits.bam O=accepted_hits_nodup.bam VALIDATION_STRINGENCY=SILENT REMOVE_DUPLICATES=true M=output.metrics

Any idea on why this behaviour persist and any solutions on this matter?
Tagged:

Answers

Sign In or Register to comment.