Science and Technology Daily, Beijing, November 5th (Reporter Zhang Mengran) - British company "Deep Thinking" has recently announced a new generation of "Alpha Fold late", which not only significantly improves accuracy, but also extends the prediction range from proteins to other biological molecules, including ligands. This model can already predict almost all molecules in the protein database (PDB), with prediction accuracy reaching atomic level. "Deep Thinking" said that the expanded functions and performance of this model will accelerate breakthroughs in the biomedical field, promote mankind to the next "digital biology" era, and provide unprecedented insights into functional research of disease pathways, genomics, bio renewable materials, plant immunity, potential therapeutic targets, drug design mechanisms, protein engineering, synthetic biology and other fields, And created a brand new platform. This time, researchers demonstrated the outstanding ability of the new generation model in predicting precise structures beyond protein folding, which can make highly accurate structural predictions for ligands, proteins, nucleic acids, and post translational modifications. This means that it will perform outstandingly in the field of antibody binding prediction; In terms of protein structure issues related to drug discovery, it clearly surpasses "predecessors" and industry standards; This model also has the ability to jointly model all atomic positions, which can more comprehensively reveal the flexibility of protein and nucleic acid interactions with other molecules. Traditional methods require the use of rigid protein structures and docking methods to predict protein ligand structures, while new generation models do not require these prior information and instead exhibit higher accuracy. It can be said that it redefines the criteria for predicting protein ligand structure, allowing previously unknown proteins to be predicted as well. The new generation of models will be applied to the design of therapeutic drugs, helping to quickly and accurately describe various types of macromolecular structures that are crucial for treating diseases. In addition, by simulating and unlocking proteins, ligands, nucleic acids, and post translational modified structures, this model can provide a faster and more accurate tool for basic biological research. 【 Editor in Chief Circle 】 Each protein in the human body contains dozens to hundreds of amino acids, and the order of amino acids determines their interactions, endowing proteins with complex three-dimensional shapes and ultimately determining their functions. For decades, scientists have used experimental techniques such as X-ray crystallography or low-temperature electron microscopy to decipher the three-dimensional structure of proteins, but this method may take months or even years and may not necessarily be effective. The emergence of 'alpha folding' is believed to have changed the rules of the game and achieved fundamental breakthroughs. At the same time, we are pleased to see that this AI is constantly upgrading and expanding, covering more complex objects with higher accuracy. (New News Agency)