Facilitating structure-based drug design for human albumin by Artificial Intelligence

With the increase of drug research and development activities, the demand for resources for clinical medical experimental research grows year by year, and the traditional biological experimental methods can no longer meet the requirements of saving the existing resources, so this paper proposes a new method combining protein drug structure prediction and AI artificial intelligence to perform structure-based drug design on the computer. Considering that the medical community has not yet obtained the complete three-dimensional structure of human albumin, it is taken as a full-text research object for subsequent analysis and research. Firstly, the standard amino acid sequence of HSA was extracted from the uniprot protein database and imported into alphafold2 for structural prediction with atomic accuracy, and then the prediction results were fed into the plip tool for non-covalent bonding interactions to explore the potential binding sites and drug pockets. The experimental results indicate that human serum albumin has high drug activity and is a suitable choice as a ligand drug, which needs to be further investigated by related scholars. This study provides a novel method to guide the drug development, which is expected to play an important role in future protein applications and drug repositioning.