After searching in this user group and analyzing the output of picard MarkDuplicates, it seems that picard does not output which of the reads it determined to be optical duplicates. That is, picard gives the number of optical and total duplicates in their metric file, but it only marks in its output bam file which reads are duplicates (optical or otherwise), and not which of those are optical duplicates.
I was requested to find a way to remove the optical duplicates and the only way I can think of doing this is to change the picard source code itself to output such a flag into the bam file (which could then be used to easily filter on). Before embarking on this task however, I'd like to know if someone has already done this and would this person be so kind as to share the code to spare me the trouble?
Thanks a lot!!!