--all is a
reads.bam file, please see the
reads.bam FAQ for more info!
With the special option
--all, ccs generates one representative sequence per ZMW, irrespective of quality and passes. For this,
--min-passes 0 --min-rq 0 --max-length 0 are set implicitly and can’t be changed; the maximum draft length filter is deactivated by this. Filtering has to be performed downstream.
The workflow of ccs with
--all changes as follows.
Exception 1: There is special behavior for low-pass ZMWs. If a ZMW has fewer than 2 full-length subreads, use the subread of median length as representative consensus, optionally with its kinetic information as forward orientation using
--all-kinetics, and do not polish.
Exception 2: Only polish ZMWs with at least two full-length subreads mapping back to the draft. Otherwise, set predicted accuracy
rq tag to
-1 to indicate that the predicted accuracy was not calculated and populate per-base QVs with
+ (QV10) the approximate raw accuracy. Kinetic information are not available for unpolished drafts.
Exception 3: Instead of using an unpolished draft without kinetic information as representative consensus sequence, if
--subread-fallback is used, fall back to a representative subread with kinetic information.
--min-passes 0 --min-rq 0 is the brute force combination that will polish every ZMW, even those that only have one partial subread, with polishing makes no difference. In contrast,
--all is a bit smarter and will only polish ZMWs with at least one full-length subread and one additional partial subread; please see previous paragraph how skipped ZMWs are represented in the output.