Google's artificial intelligence company "Deep Thinking" launched a tool on the 19th that can predict whether gene mutations are harmful to the human body, which may provide assistance for research on rare diseases. The relevant research is published in the new issue of the American journal Science. It is reported that a normal person's genome carries about 9000 missense mutations, which are mutations that cause errors in a certain amino acid of the protein. Pathogenic missense mutations can severely disrupt protein function, reduce biological adaptability, cause diseases such as cystic fibrosis or cancer, or damage brain development, while benign missense mutations are harmless. So far, about 4 million missense mutations have been observed in the human body, of which only 2% have been classified and classified as pathogenic or benign mutations. Classifying the remaining unknown missense mutations is an important challenge in the field of human genetics. The inability to accurately predict more such mutations also limits the diagnosis of rare diseases and the development and application of targeted therapies. Compared to many existing similar tools, Google's tool called "Alpha Misinterpretation" demonstrates superior capabilities. It utilizes a protein sequence database and variation structure background to identify pathogenic missense mutations. There are approximately 71 million missense mutations that may occur in human proteins, and "alpha missense" evaluates all mutations and scores the pathogenicity risk of each mutation. Based on the score, 89% of missense mutations were predicted, including 57% classified as benign, 32% classified as pathogenic, and only the remaining 11% classified as uncertain. This is far higher than the current classification that humans can determine. Koli, Vice President of Research at Deep Thinking, stated that these findings further acknowledge the impact of artificial intelligence on natural science (development). (New News Agency)