Separate sequence findings from laboratory policy
Why host preferences, deterministic findings, and laboratory SOP decisions need separate versioned records.
One detector should not silently define every laboratory's policy
A sequence-review rule answers a narrow question: what feature was found, where was it found, and by which detector version? A laboratory SOP answers a different question: must that finding block a construct, trigger a warning, or be retained for review? Combining both decisions in one hard-coded rule makes results harder to reproduce and makes local policy look like universal biology.
FactorForge's dual-profile direction keeps these responsibilities separate:
- A host profile describes codon and composition preferences for the selected context.
- A detector emits a structured sequence finding without changing its meaning by laboratory.
- A versioned SOP profile resolves that finding as required, preferred, informational, or ignored.
- The exported record preserves the profile identity, version, hash, and resolved policy.
Review is part of provenance
The same separation matters in prospective experiments. The sequence generated by an engine and the sequence eventually built should remain distinct identities. If an expert requests synonymous remediation after an advisory scan, that change is not noise and should not be attributed back to the original algorithm. It is a review event linking the finding, reviewer decision, pre/post sequence hashes, and built construct.
Applying the same frozen review policy to every comparison arm is essential. If one arm receives extra remediation, a later outcome may reflect unequal human intervention rather than the computational method.
What the current software does - and does not claim
The web workflow provides a conservative editable SOP template that can be downloaded, customized, uploaded, reset, and retained in the user's browser. It is a starting point, not a complete or approved SOP for any named laboratory. The profile configures computational design and review behavior; it does not establish synthesis acceptance, expression, yield, or wet-lab performance.