AsianScientist (Jun. 24, 2019) – A team of scientists in China has applied artificial intelligence (AI) technology to predict the structural properties of proteins. Their findings are published in the Proceedings of the National Academy of Sciences.
Living organisms depend on proteins to live and function. To better understand proteins, researchers often try to decipher protein structure. Techniques like X-ray crystallography and nuclear magnetic resonance are useful tools to probe protein structure, but some proteins are complex and harbor changeable domains, making it difficult to obtain clear spectra.
In the present study, researchers led by Professor Jiang Jun at the University of Science and Technology of China (USTC), in collaboration with US scientists, used artificial intelligence to better piece together the spectral fingerprints of proteins. The team first obtained 50,000 groups of peptide bond model molecules with different configurations using molecular dynamics simulation and quantum chemistry calculations.
The scientists next selected bond length, bond angle, dihedral angle and charge information as descriptors of protein structure, which allowed them to establish the structure-property relationship of a peptide bond in both its ground and excited state. This information was then used to build a machine learning model to predict the ultraviolet absorption spectra of the peptide bonds.
According to the authors of the study, this is the first time that AI technology has been used to perform theoretical calculation and prediction of protein spectroscopy. The study establishes the feasibility and advantages of machine learning for simulating the ultraviolet absorption spectra of the protein peptide bond skeleton, which could be useful in the interpretation of optical fingerprints of proteins.
The article can be found at: Ye et al. (2019) A Neural Network Protocol for Electronic Excitations of N-methylacetamide.
———
Source: Chinese Academy of Sciences; Photo: Shutterstock.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.
