According to a recent report on the website of New Scientist magazine, deep thinking, a British based artificial intelligence company, announced that it would publish the structure of more than 200 million proteins. In just 18 months, the company predicted the structure of almost all the proteins cataloged so far with the "alpha folding" algorithm, and solved one of the most important challenges in the field of biology. It will help cope with antibiotic resistance, accelerate drug development and completely change basic science. For decades, the determination of protein shape based on amino acid sequence has been a major problem in the field of biology. At the end of 2020, "deep thinking" announced that the company's "alpha folding" algorithm can accurately predict the structure of folded proteins; By 2021, the artificial intelligence has been able to map 98.5% of the protein in the human body. Recently, the company announced that it would publish the structure of more than 200 million proteins, almost all of which were incorporated into the globally recognized protein research library UniProt. "Deep thinking" is also cooperating with the European Institute of bioinformatics under the European Molecular Biology Laboratory to create a searchable database "alpha folded protein structure database". Researchers can easily and freely access relevant information, making searching for protein structure almost as simple as web search tools. Many scientists are using "alpha folding" to promote research in many fields. For example, Matt Higgins of Oxford University and others are studying a key protein that they believe is the key to interrupt the life cycle of malaria parasites, hoping to develop an effective malaria vaccine; Some scientists have also designed new enzymes to decompose plastic waste and further understand the proteins that make bacteria resistant to antibiotics. Keith Williamson of Imperial College London said that "alpha folding" has changed biological research, but there are still some problems. For example, it cannot extract any amino acid sequence and accurately simulate their folding mode, nor can it reveal the complex interactions between proteins. In addition, its accuracy needs to be improved. Deep thinking said it was working to improve the accuracy of the tool to further understand how proteins are generated and how cells work. (news agency)