Background
I'm the maintainer of atropos and am winding it down in favor of actively-maintained tools like fastp. A capability atropos has that fastp doesn't, which I'd like to propose here, is named presets for common library types.
Proposal
Add a --preset <name> CLI flag (or similar) that expands into a documented bundle of trimming/adapter settings appropriate for a specific library preparation. Users who know their library type but don't know which individual fastp knobs to set would get defaults that match published best practices.
Suggested presets (candidates; final list up to maintainers)
- Bisulfite-RRBS — clip the
CGG end-repair motif, suppress poly-X for C/T-biased reads (see Bismark docs).
- Bisulfite-WGBS (directional) — standard R1/R2 end clipping.
- Bisulfite-non-directional / Bisulfite-Swift / TruSeq-Epignome — tool-specific end clips from the respective kit documentation.
- Small-RNA / miRNA — Illumina small-RNA 3′ adapter, short min length.
- TruSeq and Nextera — vendor-specific default adapters (fastp already has these in
knownadapters.h, this just surfaces them as a preset).
- Amplicon — aggressive overlap merge, strict quality filtering.
Why this is useful
Right now fastp has great defaults for generic Illumina shotgun data but users running bisulfite, small-RNA, or vendor-specific workflows have to assemble the correct combination of individual flags themselves, often from scattered documentation. A preset tells them "I know my library type; do the right thing" and reduces trimming-related bugs in downstream analyses.
Prior art
Happy to supply the atropos cut-rule tables as a starting point if useful.
Background
I'm the maintainer of atropos and am winding it down in favor of actively-maintained tools like fastp. A capability atropos has that fastp doesn't, which I'd like to propose here, is named presets for common library types.
Proposal
Add a
--preset <name>CLI flag (or similar) that expands into a documented bundle of trimming/adapter settings appropriate for a specific library preparation. Users who know their library type but don't know which individual fastp knobs to set would get defaults that match published best practices.Suggested presets (candidates; final list up to maintainers)
CGGend-repair motif, suppress poly-X for C/T-biased reads (see Bismark docs).knownadapters.h, this just surfaces them as a preset).Why this is useful
Right now fastp has great defaults for generic Illumina shotgun data but users running bisulfite, small-RNA, or vendor-specific workflows have to assemble the correct combination of individual flags themselves, often from scattered documentation. A preset tells them "I know my library type; do the right thing" and reduces trimming-related bugs in downstream analyses.
Prior art
--rrbsand--non_directionalflags are another reference.Happy to supply the atropos cut-rule tables as a starting point if useful.