Elucidating Structures of Complex Organic Compounds Using a Machine Learning Model Based on the ¹³C NMR Chemical...

Area of Science:

• Computational Chemistry
• Organic Chemistry
• Machine Learning

Background:

• Accurate structural and stereochemical assignment of complex organic molecules is crucial in chemistry.
• Existing methods may face challenges in handling complex structures and verifying assignments efficiently.

Purpose of the Study:

• To introduce a novel computational protocol (SVM-M) for reliable structural and stereochemical assignments.
• To demonstrate the protocol's capability in simultaneously classifying correct structures and comparing candidate structures.

Main Methods:

• Combining Support Vector Machine (SVM) model with accurate 13C chemical shift calculations.
• Utilizing the xOPBE/6-311+G-(2d,p) level of theory for chemical shift computations.
• Leveraging the dual role of SVM decision values for classification and comparison tasks.

Main Results:

• The SVM-M protocol achieved a success rate of approximately 100% on a dataset of 760 molecules (15,928 13C chemical shifts).
• Demonstrated high confidence in identifying structural and stereochemical assignments for complex organic compounds.
• Successfully handled both classification (correct/incorrect structure) and comparison (most likely structure) problems.

Conclusions:

• The SVM-M protocol is a versatile and powerful tool for routine structural and stereochemical assignments.
• It offers an efficient solution for detecting mis-assignments in complex organic molecules, including natural products.
• The method provides high accuracy and confidence, facilitating advancements in chemical structure elucidation.