Very fast multi-threaded sort/merge for BAM files.
Packed with features:
- Reduced run times from multi-threading and by combining sort,merge & index creation in one step.
- Mark duplicates while you sort.
- Option to included strand as part of the sort key.
- Picard like handling of @PG and @RG identifiers
- Uses a stable sort/merge algorithm that will not change the order of alignments with the same sort key.
- Option to add or replace @RG record
- Smart handling of @SQ records means order does not have to be the same in files being merged.
- Creates BAM index file.
“Re-sorting aligned files by read position … is a common task. The fastest algorithm to our knowledge is Novosort … This algorithm is so efficient that it saturated the peak node injection bandwidth on Blue Waters (9.6GB/s)”
Mainzer et al, Instrumenting Human Variant Calling Workflow, Blue Waters 2015 Annual Report