TY  - JOUR
T1  - How Much Structure Has the Spectrum Actually Proven? A Confidence-Graded Workflow for Combining HRMS, NMR, Chiroptical Data, and Biosynthetic Plausibility
A1  - Mikko Lahtinen
A1  - Elina Salo
A1  - Juhani Virtanen
JF  - Specialty Journal of Pharmacognosy, Phytochemistry, and Biotechnology
JO  - Spec J Pharmacogn Phytochem Biotechnol
SN  - 3062-441X
Y1  - 2024
VL  - 4
IS  - 2
DO  - 10.51847/NeG6SB9yyV
SP  - 40
EP  - 49
N2  - Natural-product structures are commonly communicated as finished molecular identities, yet the analytical evidence underlying a drawing is heterogeneous in what it can establish. High-resolution mass spectrometry constrains elemental composition and fragmentation-compatible substructures; NMR defines connectivity and, with suitable restraints or calculations, stereochemical relationships; chiroptical methods can support absolute configuration; and biosynthetic reasoning can test chemical coherence. None of these evidential roles is interchangeable. This article develops an original, proposed analytical confidence scheme that treats structure elucidation as a sequence of claim-specific inferences rather than a binary identified/not-identified decision. The scheme separates observed data, candidate generation, candidate discrimination, stereochemical assignment, plausibility assessment, conflict detection, and reporting. It also distinguishes incomplete evidence from contradictory evidence: missing support should lower resolution, whereas material incompatibility should reopen candidate space or trigger escalation to more discriminating measurements. The central proposal is that confidence should attach to the structural proposition actually tested—composition, constitution, relative configuration, absolute configuration, or full molecular assignment—rather than to the molecule as a whole. Orthogonality is therefore judged by whether a new modality contributes an independently constraining observable, not merely another score derived from the same assumptions. The framework is intended to improve transparency in natural-product structure assignment and revision, but it is not a validated probability model and does not supply universal numerical thresholds. Its applicability remains limited by sample availability, spectral quality, conformational complexity, candidate-space completeness, computational domain, and the feasibility of orthogonal validation.
UR  - https://galaxypub.co/article/how-much-structure-has-the-spectrum-actually-proven-a-confidence-graded-workflow-for-combining-hrms-ejrdgnvppjhqurz
ER  -