Core Reading Recently, six departments including the Ministry of Science and Technology issued a document, focusing on creating a number of major scenarios and expanding the application of artificial intelligence. High level scientific research activities are one of them. Nowadays, China's AI technology is developing rapidly, and has advantages in data acquisition, experiment prediction, result analysis, etc. Life science, mathematics, chemistry, space science and other disciplines have embraced AI. The rich application scenarios also feed back the development of technology and promote the upgrading of industrial intelligence. From daily life to scientific research, today, China's artificial intelligence technology is developing rapidly, and data and computing resources are increasingly rich. Application demand is an important driving force for technological progress, and new technologies are often constantly improved and mature in "use". In order to promote the implementation of artificial intelligence, the Ministry of Science and Technology and other six departments jointly issued the Guiding Opinions on Accelerating Scene Innovation and Promoting High Quality Economic Development with High Level Application of Artificial Intelligence a few days ago, focusing on creating a number of major scenes and expanding the application of artificial intelligence, including high-level scientific research activities. As an enabling means, how can AI bring new research methods and inject "intellectual momentum" into economic development? It is closely integrated to help scientific research become more efficient and accurate. Among many disciplines, life science research and artificial intelligence are closely integrated. One of the hot directions is to predict protein structure. Protein has a three-dimensional structure, and its primary structure (sequence) is composed of multiple amino acids in series. The three-dimensional structure determines the function of proteins in cells. Many diseases are caused by the structural abnormalities of important proteins in the body. Therefore, only by drawing a "three-dimensional map" of important proteins in the human body can we find out the targets of drugs acting on the human body and develop accurate and effective new drugs. Traditionally, scientists used cryoelectron microscopy, X-ray, nuclear magnetic resonance and other methods to observe the three-dimensional structure of proteins, but this process is time-consuming and expensive. "Taking the cryoelectron microscope as an example, it will cost tens of millions of yuan to arrange an observation platform, and it will take a long time for researchers to draw the protein structure," said He Jingzhou, director of Baidu PaddlePaddle Propeller Biocomputing Platform. Due to the high difficulty, long experiment period and high cost, the number of three-dimensional protein structures observed by traditional methods is very limited up to now. In contrast, amino acid sequencing is much easier. Why can't we predict the structure of a protein based on its amino acid sequence? As early as 1972, American biochemist Christian Amphenson put forward this idea in his Nobel Prize winning speech. Accurate prediction of the three-dimensional structure of proteins from the primary structure is what AI is good at. However, human efforts to analyze proteome have been slow. It is explained that, on the one hand, due to the small amount and low quality of existing biological data, there is a lack of sufficient samples for in-depth learning; On the other hand, due to the maturity of AI algorithms, processes are also required. In recent years, with the dramatic increase of biological data and the optimization of artificial intelligence technology, scientists have established more accurate prediction models. In December 2020, in a competition, the AI program "Alpha Folding" shines brightly, and its predicted results are similar to most experimental data. This proves that the prediction of protein structure, artificial intelligence has been