Dos Santos Fernando Martins, da Silva Mota Gunar Vingre, Martorano Lucas Haidar, de Albuquerque Ana Carolina Ferreira, da Silva Claudinei Alves, da Silva Adalberto Manoel, de Jesus Chaves Neto Antônio Maia, Valverde Alessandra Leda, Cardoso Evani Ferreira, Costa Fabio Luiz Paranhos
Department of Organic Chemistry, UFF, Niterói, Brazil.
Natural Science Faculty, ICEN, UFPA, Belém, Brazil.
Magn Reson Chem. 2022 Jun;60(6):533-540. doi: 10.1002/mrc.5261. Epub 2022 Mar 9.
The combination of computational methods and experimental data from Nuclear Magnetic Resonance (NMR) is a considerably valuable tool in the elucidation of new natural product structures and, also, in the structural revision of previously reported compounds. Until recently, only classical statistical parameters were used, for example, linear correlation coefficient (R ), mean absolute error (MAE), or root mean square deviation (RMSD), as a way to statistically "validate" the structure pointed out by experimental NMR spectra. Regarding the resolution of the relative configuration of organic molecules, novel tools were available in the last few years to assist in the NMR elucidation process. The most relevant are DP4+, which is based on a Bayesian probability, and ANN-PRA, which is based on artificial neural networks. The combined application of these tools has become the most accurate and important alternative to solve structural and stereochemical problems in natural product chemistry. Therefore, herein, in this case study, we intended to promote these novel tools, exploring the strengths and limitations of each approach in resolving the relative configuration of the sesquiterpene alpha-bisabol. We also highlighted the advantages of the complementary use of H- and C-DP4+ to obtain optimal results in the differentiation of the stereoisomers, validating the proposal with ANN-PRA method.
计算方法与核磁共振（NMR）实验数据相结合，是阐明新天然产物结构以及对先前报道化合物进行结构修正的非常有价值的工具。直到最近，仅使用经典统计参数，例如线性相关系数（R）、平均绝对误差（MAE）或均方根偏差（RMSD），作为从统计学角度“验证”实验NMR光谱指出的结构的一种方式。关于有机分子相对构型的解析，在过去几年中有一些新工具可用于辅助NMR解析过程。其中最相关的是基于贝叶斯概率的DP4 +和基于人工神经网络的ANN - PRA。这些工具的联合应用已成为解决天然产物化学中结构和立体化学问题的最准确且重要的方法。因此，在本案例研究中，我们旨在推广这些新工具，探索每种方法在解析倍半萜α-红没药烯相对构型方面的优势和局限性。我们还强调了互补使用H - 和C - DP4 +在区分立体异构体方面获得最佳结果的优势，并使用ANN - PRA方法验证了该提议。
