Question: How can I identify PCR duplicates from Long Ranger output?
Answer: PCR duplicates are marked in the BAM file, but not filtered. To identify PCR duplicates you will need to interpret the flags in the second column of the BAM file, or more appropriately the second column of the non-binary equivalent SAM file.
The bit flag that corresponds to PCR or optical duplicate is 1024. However, an entry with a flag other than 1024 may also be a PCR or optical duplicate. For example a bit flag of 1091 would be a paired read (1), mapped in proper pair (2), first in pair (64), and an optical duplicate (1024) you add all of the bit flags together to get 1091.
For more information please see the SAM/BAM format specification.
Note: This article applies only to BAM files generated by Long Ranger using linked reads and is not relevant for gene expression assays that use UMIs to identify duplicate reads. For more information on UMI based deduplication please see the article here.