I need to re-map two kinds of BAMS: a) the re-calibrated BAMs and b) hg19 mapped BAMs to hg38.
I am looking into the GATK tutorial: https://software.broadinstitute.org/gatk/guide/topic?name=tutorials#optionB, on "Convert aligned BAM to uBAM and discard problematic records using RevertSam", and below is an example.
java -Xmx8G -jar /path/picard.jar RevertSam \
I=6484_snippet.bam \
O=6484_snippet_revertsam.bam \
SANITIZE=true \
MAX_DISCARD_FRACTION=0.005 \ #informational; does not affect processing
ATTRIBUTE_TO_CLEAR=XT \
ATTRIBUTE_TO_CLEAR=XN \
ATTRIBUTE_TO_CLEAR=AS \ #Picard release of 9/2015 clears AS by default
ATTRIBUTE_TO_CLEAR=OC \
ATTRIBUTE_TO_CLEAR=OP \
SORT_ORDER=queryname \ #default
RESTORE_ORIGINAL_QUALITIES=true \ #default
REMOVE_DUPLICATE_INFORMATION=true \ #default
REMOVE_ALIGNMENT_INFORMATION=true #default
I have few questions:
1) The re-calibrated BAMs are generated using different GATK versions (3.4-3.6). What version of picard should we use for RevertSam?
2) Can we also use RevertSam on the mapped BAMs? If so, what parameters should be used?
3) It would be much appreciated if you can give more information on the parameters for this RevertSam:
a) MAX_DISCARD_FRACTION=0.005 --> if we don't want to discard any reads, we should set this value to zero, right?
b) What's your recommendation on ATTRIBUTE_TO_CLEAR? It's not clear to me how this parameter will work on top of REMOVE_ALIGNMENT_INFORMATION.
Thanks!