Pose-recovery-qualified multi-conformer, multi-seed docking prioritizes GalA2 as a structural hypothesis for human N-terminal maltase-glucoamylase

Defined oligogalacturonides have been reported to inhibit yeast -glucosidase, but direct evidence against human N-terminal maltase-glucoamylase (NtMGAM) is lacking. We asked whether idealized GalA2 and GalA3 generate reproducible geometries in the acarbose-defined orthosteric site. Chain A of human NtMGAM structure 2QMJ was treated as a rigid, dry receptor. Vina-Carb 1.0 was the prespecified primary method and conventional AutoDock Vina 1.2.7 a diagnostic method used only to assess method dependence. Two representative conformers (cluster medoids) from short ligand-only explicit-solvent simulations per ligand were docked across ten fixed seeds. Top-five poses were subjected to prespecified active-site, clash, integrity, carbohydrate-torsion, recurrence, and cross-conformer criteria. Acarbose pose recovery qualified the search procedure. A qualifying acarbose pose was recovered in all ten Vina-Carb seeds within ranks ; every representative occurred at rank 2, with heavy-atom RMSD of and of 18 crystallographic contacts recovered. Across 80 jobs, 1,590 poses were generated and 400 top-five poses analyzed. Vina-Carb retained 27/100 GalA2 poses, which formed one recurrent family spanning both conformers and all ten seeds, meeting the geometric and cross-conformer recurrence criteria. For GalA3, 29/100 poses qualified and formed three families spanning seeds, but none crossed conformers, so the cross-conformer criterion was not met. Diagnostic Vina produced cross-conformer GalA3 families, indicating method dependence. GalA2 is prioritized as a reproducible orthosteric structural hypothesis for experimental testing, whereas GalA3 remains unresolved. These calculations do not demonstrate binding, inhibition, altered starch digestion, or glycemic effects; human-enzyme kinetics and orthogonal structural or biophysical evidence are required.