Quantitative Structure-Activity Relationships in Toxicity and Binding Affinity Prediction of Anti-Cancer Drug-Like Molecules Utilizing Neural Networks Techniques

Background and aims: Quantitative structure-activity relationships (QSAR) is an approach topredict small molecule properties based on their physical and biological characteristics. Traditionally,older QSAR methods were complicated and time-consuming, but artificial intelligencebrings new and better solutions to QSAR modeling and drug discovery. It also consists of severalmethods and analyses of chemical molecules in the vast majority of disciplines, which has alsobeen highlighted in medicinal chemistry. Furthermore, its critical role will be displayed more andmore in the treatment of cancer with personalized medicine. One of the most significant challengesin oncology is the identifying potential drug candidates that can effectively target cancer cellswhile minimizing toxicity to healthy cells. the combination of novel QSAR and neural networkshas emerged as a promising approach in anti-cancer drug discovery and development.Method: The part of the updated review was applied to the study of standard computational differentAlgorithms in several documents. The Scopus and Web of Science databases were scannedin multiple areas, including pharmacology, pharmaceutics, toxicology, and biochemistry. A bunchof studies were obtained in range of ۲۰۲۱ to ۲۰۲۲. Review articles, conference papers, and similarpapers were excluded. Ultimately, ۱۷ recent articles were screened as eligible for this study andincluded in the review. This study compares several original research methods to address problemsand discusses the accuracy and precision of each algorithm in order to deal specific issues.Results: Deep learning techniques, like artificial neural networks, can predict interplays and bindingaffinity of drugs and targets by using the atomic coordinates of protein-ligand interactions incomplex structures. Screening large datasets of candidate molecule’s physiochemical behaviors,to facilitate new drug discovery process. For instance, the identification of novel inhibitors of theprotein kinase CK۲. CK۲ has been shown as a likely target for cancer therapy, due to its involvementin several signaling pathways that are dysregulated in cancer cells. However, innovativein vitro and in vivo assays are being recruited parallel to drug discovery. Intelligent algorithmsintegrated into QSAR provide more accurate and efficient information about drug-candidate moleculespharmacokinetics, pharmacodynamics, and safety profiles.Conclusion: Artificial intelligence is expected to play a significant role in developing new drugmolecules due to the rapid development of cutting-edge technologies and cost-effectiveness models.Also, make researchers and pharmacists potent during the drug development procedure. In addition,computational approaches can perform well in the virtual execution of complicated cancermodels treatments with target cell receptors or molecules. moreover, deep learning techniques areenhancing the progress of drug repurposing (finding new therapeutic uses for existing drugs) atpresent. The combination of QSAR and artificial intelligence has revolutionized and transformedour thinking in the direction of a new paradigm for drug design. On the other hand, it will becompatible with sustainable development policies.

کلیدواژه ها:

deep learning ، QSAR ، data mining ، oncology ، drug discovery