Reaching computational predictions of spectroscopic data, like nuclear magnetic resonance (NMR), infrared spectroscopy (IR), and mass, has become a commonly utilized method for aiding in the prediction of glycan structure. Nevertheless, traditional quantum mechanics simulations still demand significant effort. CD BioGlyco develops machine learning methods to correlate experimental and computational NMR data to facilitate Analysis/Prediction Service. Therefore, we provide our clients with thorough Glycoinformatics-assisted Structural and Functional Prediction Services, which encompass NMR-based, IR-based, and Mass-based Glycan Structure Prediction Service. The details of our glycoinformatics-assisted NMR-based glycan structure prediction service are as follows:
We use glycoinformatics tools to analyze the chemical structure and properties of glycans to predict glycan structures. By using glycoinformatics tools to generate plausible glycan structures and use them as models for subsequent simulations and calculations. In addition to predicting the structure of glycan molecules, we help you understand the mechanisms by which glycans interact with other biomolecules. By modeling and simulating different types and quantities of carbohydrates, we help you better grasp the interactions between multiple biomolecules involved in complex reaction networks in organisms.
At the same time, we compare the generated glycan structure with the NMR observations and determine the optimal glycan structure through calculations and simulations. This is done by calculating the difference between the NMR observations and the model and using an optimization algorithm to adjust the glycan structure to best match the experimental data. The NMR technique provides information about the structure and conformation of glycans. Structural and conformational information about glycans is obtained by measuring the interactions between different nuclear spins.
Technology: Machine learning
Journal: Frontiers in Natural Products
IF: 4.362
Published: 2023
Results: This article focuses on a method called DU8ML, which combines machine learning and density functional theory (DFT) techniques for calculating NMR chemical shifts and spin-spin coupling constants (SSCCs) of large natural products. The DU8ML method is capable of calculating the NMR parameters with high accuracy in a short time, by using experimental data for training in machine learning. It has been used to obtain quantum mass NMR chemical shifts at a low computational cost and also to establish correlations between experimental and computational data to determine the most probable structures. The article also mentions the application of the DU8ML method in structural characterization, including the detection of erroneous atomic connections, stereochemical errors, and other problems related to structural characterization.
Fig.1 Schematic illustration of the application of machine learning in computational NMR-assisted determination of molecular structure. (Cortés, et al., 2023)
CD BioGlyco provides a comprehensive glycoinformatics-assisted NMR-based glycan structure prediction service. By combining glycoinformatics tools and NMR technology, we quickly and accurately predict the structure of complex glycans and help our clients to solve the difficulties in glycan structure analysis. If you are interested in our service, please feel free to contact us.
Reference
We envision a future where the intricate world of carbohydrate is no longer shrouded in mystery, but rather illuminated by the power of cutting-edge computational tools.