Small-Molecule 1H and 13C NMR: From Chemical-Shift Prediction to Automated Structure Elucidation

Abstract 1H and 13C NMR chemical-shift prediction has progressed from empirical rules and experimental databases to graph neural networks, three-dimensional molecular representations, and approaches that integrate quantum chemistry with machine learning; its applications have likewise expanded from chemical-shift calculation for known structures to experimental spectral assignment, candidate-structure discrimination, and automated structure elucidation. Forward structure-to-NMR prediction can generate predicted spectral-library data for candidate molecules and support structure retrieval and candidate screening by matching predicted and experimental chemical shifts; inverse NMR-to-structure analysis progressively constrains the candidate space using structural constraints encoded in experimental spectra. This Review systematically surveys the development and growing integration of these two directions, focusing on how chemical-shift prediction contributes to peak-to-atom assignment, candidate-structure comparison, and automated structure elucidation, and examining how prediction error, candidate coverage, and experimental data quality affect final structural decisions. More recently, agentic approaches have begun to organize existing NMR prediction and structure-elucidation tools into iterative analytical systems that continuously update candidate structures from current spectroscopic evidence and select subsequent computational or experimental operations; further progress in automated structure elucidation depends on reliable predicted spectral libraries, clearly defined model applicability domains, and structural conclusions that can be tested against experimental spectroscopic evidence.