Duplicates are defined as having multiple templates whose aligned 5’ coordinates match. For a paired-end template this requires both primary reads to have matching 5’ coordinates. Coordinates are based on the unclipped position of the read against the reference. Reads also need to match orientation. When a duplicate is detected, the overall highest quality template is kept and all others have the duplicate flag set.
For primary reads, this definition is the same as used in Picard (v2.10.3) and Biobambam2 (bamstreamingmarkduplicates v2.0.57). None of these tools use supplementary data when determining a duplicate. However unlike Picard, supplementary reads in a duplicate template do not have their flags modified in Samtools by default.
Copyright © 2019 Genome Research Limited (reg no. 2742969) is a charity registered in England with number 1021457. Terms and conditions.