Precursor isotope envelopes currently ignore modification C/S contributions and choose a mass-estimated fallback independently for each entry.
Agreed policy, implemented in draft #145:
- Keep the existing fast three-bin C/S distribution; add signed modification C/S contributions. Other elements remain outside the approximation.
- Resolve one method for the library. If any stored target or shipped decoy lacks usable counts, use mass-estimated C/S for every entry.
- Synthetic mass-shift decoys retain parent composition/envelope. Mass estimation uses the unshifted parent mass. Shipped decoys use their own stored facts under the same library-wide policy.
- Independently supplied formulas with an explicit neutral/observed-ion basis can establish C/S without a sequence. Check comparable C/S conflicts; report unsupported chemistry and numerical limits.
- Cache envelopes during library finalization and report method/coverage/reasons through the shared CLI/viewer plan and Parquet metadata.
Example: PEPTC[UNIMOD:4]IDEK changes from C43/S1 to C45/S1. A library containing one mass-only modification instead selects mass-estimated C/S for all entries.
Slice 3 of #97. Broader sequence-free scoring and changes to the underlying chemical approximation remain separate work.
Precursor isotope envelopes currently ignore modification C/S contributions and choose a mass-estimated fallback independently for each entry.
Agreed policy, implemented in draft #145:
Example: PEPTC[UNIMOD:4]IDEK changes from C43/S1 to C45/S1. A library containing one mass-only modification instead selects mass-estimated C/S for all entries.
Slice 3 of #97. Broader sequence-free scoring and changes to the underlying chemical approximation remain separate work.