13- 17 November 2019 Shenzhen Convention & Exhibition Center
AI helps study optical properties of proteins
News sources:CHTF Organizing Committee Release Date:2019-06-12

 

AI machine learning is proved instrumental in predicting the optical properties of proteins, according to Professor Jiang Jun and his colleagues of the Hefei National Laboratory for Physical Sciences at the Microscale in China University of Science and Technology. [Ma Xuejing/China Daily] 

 

The neural network technology in AI machine learning has proved instrumental in predicting the optical properties of proteins, according to Professor Jiang Jun and his colleagues of the Hefei National Laboratory for Physical Sciences at the Microscale in China University of Science and Technology. 

  

The spectral response signals of proteins, especially the ultraviolet spectrum, can be interpreted by theoretical simulations to reveal precise protein structures, as well as provide vital information for life sciences and medical diagnosis. However, the structure of proteins is extremely complex and variable, and a large number of high-precision quantum chemical theoretical calculations are required. 

  

The researchers first screened through AI algorithms and then constructed a structure-activity relationship between the ground state structure of the peptide bond and its excited state properties through a neural network. They then predicted the ultraviolet absorption spectrum of the peptide bond. 

  

This is the first time AI technology has been used to theoretically predict the spectral study of proteins. A large amount of data was obtained through theoretical calculations and the technology can effectively promote the accurate analysis of the spectrum, as well as the discovery of protein structure after training. 

访问统计: