Three separate BacDive galactoside rows: finite structural review

2026-09-29. Related to #286 and #1249; not a global identifier replacement,
historical CAS-cohort closure, or source execution/merge acceptance.

Exact cohort and evidence

The complete original rows at physical lines 28, 120 and 271 of the current
legacy bacdive_mappings.tsv all specify 4-Nitrophenyl-alpha-D-galactoside,
CAS-RN:7493-95-0 and EC:3.2.1.22. They incorrectly supply CHEBI:54684, which
the complete selected native ChEBI scan does not declare or replace. The full
original rows and native records are already preserved in
issue286-glucoside-authority.1cVWUE/native-authority.json, SHA-256
71b06e518f68c9bc3f64aaa98a0688191787d4bb43d4007061b33ef372e68c65.
The legacy input SHA-256 is
f39e9753fe2879877b2cba51c9896f44515f3a5785e790deb0d0bd5ca02c9379.

The complete native CHEBI:546840 record has the matching explicit alpha-D-
galactopyranoside name, an alpha-D-galactopyranoside parent CHEBI:46953, and
InChIKey IFBHRQDFSNCLOZ-IIRVCBMXSA-N. The entire native InChI matches the
independent primary manufacturer's N0877 record,
including the /t8-,9+,10+,11-,12+/m1/s1 stereolayer. The glucoside in #1249
instead has 9- and a different stereospecific key. Equal formulas or similar
ChEBI digits are not the basis for the correction.

N0877's product-name and application sections support use for alpha-galactosidase
activity. Its marketing subtitle inconsistently says alpha-glucosidase; that
conflict is disclosed, not silently treated as corroboration. Official
IUBMB EC3.2.1.22 confirms the
alpha-galactosidase identity and reaction scope. Neither source establishes that
this commercial reagent was used in a particular API kit experiment or that
every enzyme assigned the EC accepts it.

The public ChEBI546840 page could not be fetched through the web tool during
this review. The ontology evidence above is the already hashed complete local
native record, not a claim of a successful live-page retrieval.

Conflicting ancillary identifier

Physical line120 also claims KEGG:C01083. The current primary
KEGG C01083 record is alpha,alpha-trehalose,
formula C12H22O11, not the nitrophenyl galactoside. This ancillary assignment
must remain an unendorsed original claim in the audit, never a corroborating
identity assertion. The reviewed active BacDive EC-substrate emitter consumes
EC_ID and CHEBI_ID, not this KEGG cell. No CAS or KEGG identity edge is licensed.

Proposed bounded disposition

Correct the CHEBI endpoint of only these three complete reviewed row identities
to CHEBI:546840, preserving source multiplicity before normal finalization,
all original claims, and existing edge relation/provenance/tier. Retain the
contradictory KEGG claim only as raw audit evidence. Do not alter the raw legacy
table; use a finite reviewed transform correction policy. Do not rewrite line119
to this galactoside or create a global 54684 replacement.

Require independent scientific/code review and immutable regression tests:
three galactoside corrections; separate glucoside correction; negative drift
and unrelated-row controls; no KEGG/CAS identity or assay-reagent inference;
unchanged organism-assay observations; audit/finalization coverage and actual
source rebuild before acceptance. Native absence does not establish official
obsolescence. No mapping, source output, or production archive was edited here.