Resistome profiling can establish, with reasonable confidence, that a resistance determinant is present in a bacterial genome; converting that observation into a statement about which antibiotic a clinician should prefer is a distinct and largely unsolved engineering problem. This paper specifies and describes the architecture of a sixstage computational framework that maps gut-bacterial resistance determinants onto the drug classes of the Antibiotic Resistance Ontology and thence onto sixteen clinically prescribed antimicrobial classes, before applying an explicit guidance rule. The framework was exercised on a deliberately small, exhaustively inspectable reference panel of ten gut-relevant bacterial strains and six well-characterised resistance determinants, with determinant calls curated from the primary literature rather than generated by sequence annotation. We describe the rationale for a reference-panel design, the provenance of every determinant call, the two-step ontology-to-clinical-class mapping, and a revised three-state decision rule — avoid on positive determinant evidence, avoid on recorded intrinsic resistance, or no evidence — that replaces an original binary rule shown elsewhere to generate clinically unsound recommendations. The architecture is designed so that every stage can be inspected, re-run and corrected independently, a property we argue is a precondition for any resistome-based decision aid intended for eventual clinical use. Companion evaluation results are reported separately.

Seshu Kumari K, Dr. Munesh Kumar Sharma, "A Six-Stage Computational Framework for Translating Gut Resistome Determinants into Antibiotic-Class Guidance: Architecture, Reference Panels and Decision Logic" International Journal of Scientific Research and Engineering Development, V9(5): Page(361-368) September - October 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.